ISCB Communities of Special Interest (COSIs) are member communities of shared interest that have self-organized and have multiple activities or interactions throughout the year, rather than solely meeting during the ISMB conference in the COSI track. An important goal of any COSI is to foster a topically-focused collaborative community wherein scientists communicate with one another on research problems and/or opportunities in specific areas of computational biology. Such communication is often in the form of meetings, but can also be through other social media tools that allow for vibrant participation in a virtual environment. If you are interested in starting an ISCB COSI please review the COSI Guidelines and send your exploratory questions to Diane E. Kovats, ISCB Chief Executive Officer ( dkovats@iscb.org)
The 3D Special Interest Group (SIG) COSI focuses on all aspects of molecular biology that can be informed by a computational perspective. In broad terms this includes structural bioinformatics approaches dealing with the computational analysis of biological processes at the molecular level from a structural perspective as well as the simulation of biological processes from a molecular biophysics’ perspective. Themes relevant for 3DSIG include, but are not restricted to: Prediction of Biomolecular structure and function, Dynamics of biomolecules and its role in biomolecular function, molecular bioisosterism (detection of similarities) and its role in biological function, biomolecular docking, drug design, 3D quantitative structure activity relationships (3D-QSAR), structurally informed cell-biology, etc.
Leadership Structure
Rafael Najmanovich, rafael.najmanovich@umontreal.ca, University of Montreal, PI CA Co-chair
Douglas Pires, douglas.pires@unimelb.edu.au, University of Melbourne, PI AU Co-chair
Gonzalo Parra, parra.gonzalo@gmail.com, Barcelona Supercomputing center, ES Co-chair
Alexey Porollo, Alexey.Porollo@cchmc.org, Cincinnati Children's Hospital Medical Center,
PI US Track, Editor
Anna Marabotti, amarabotti@unisa.it, University of Salerno, PI IT Track, Editor
Ray Arjun, rayarjun@gmail.com, Indraprastha Institute of Information Technology (IIIT), PI IN Track, Editor
Javier Alfaro, javier.alfaro@proteogenomics.ca, International Center for Cancer Vaccine Science (University of Gdansk), PI PL/ Track, Editor
Lam Su Datt, sudatt@ukm.edu.my, Universiti Kebangsaan Malaysia, PI Malaysia Track, Editor and Webinar Coordinator
Stephen Yi, Stephen.yi@austin.utexas.edu, The University of Texas at Austin, PI US Webinar Coordinator
Daniel Buchan, daniel.buchan@ucl.ac.uk, UCL PI UK Track, Editor and Training Coordinator
Rafaela Ferreira, rafaelasalgadoferreira@gmail.com, Universidade Federal de Minas Gerais, PI Brazil Track, Editor
Chris Kieslich, kieslich@auburn.edu, Auburn University, PI US Track, Editor and Training
Coordinator
Stuart MacGowan, s.macgowan@dundee.ac.uk, University of Dundee, PI UK Track, Editor and
Comms
S. Byamal sbyamal@gmail.com, University of Kansas, Postdoc US Track, Editor
Ravinder Abrol, ravinder.abrol@csun.edu, California State University, PI US Track Editor and Hackaton Coordinator
Carlos Rodrigues, carlosmr12@gmail.com, University of Queensland, Postdoc AU Training
Coordinator and Hackaton Coordinator + Comms
Sara Capponi, sara.capponi@ibm.com, IBM Research NA US Track, Editor
Bruna Moreira, munamomo@gmail.com, University of Melbourne, PhD Student, AU Training Coordinator+ Comms
Sandro Izidoro, sandroizidoro@unifei.edu.br, Universidade Federal de Itajubá, PI Brazil Track
Editor
Raul Secchi, raul.secchi@nbi.ku.dk, University of Copenhagen, Postdoc Denmark Track Editor
Activities Over the Last Year
3DSIG is an ISMB Community of Special Interest (COSI) that covers the latest and most innovative research in the all aspects of molecular biology and biophysics informed by structural data through computational methods.
3DSIG was formed in 2004 by Ilan Samish and John Moult initially as a satellite meeting. In 2006 3DSIG had its second installment and with the inclusion of Rafael Najmanovich to the organizing committee in 2006, it became an annual event. Over the years 3DSIG has had a broad selection of keynote speakers.
ISMB was held in Lyon, France. This time the 3DSig keynote lectures were delivered by Prof. Burkhard Rost (Technical University of Munich, Germany, TUM) and Prof. Arne Elofson (Karolinska Institute, Stockholm, Sweden).
The same procedure as for 3DSig 2022 was followed to fill the 2 days programme with selected talks.
In addition to the ISCB Academy webinars, 3DSig has started a collaborative webinar series jointly organized with the 3DBioinfo Community from Elixir Europe.
Activities outside of ISMB:
- ISCB Academy webinar: Dr. Pedro Beltrao “Clustering predicted structures at the scale of the known protein universe”. October 5, 2023 at 10:00 AM EDT.
- ISCB Academy webinar: Dr. Mohd Firdaus Raih “Substructural similarity searching: having more fun(ctions) with increasing biological structure data”. February 14, 2023.
Since 2023, 3DSig collaborates with the 3DBioinfo Community from Elixir and the ISCB Student Council to organize a webinar series focused on different Structural Bioinformatics topics. The webinars are mainly organized by Dr. Gonzalo Parra and Dr. Neeladri Sen.
These webinars were organized since April:
April 18th, 17 hs CET:
Dr. Merveille Eguida (University of Strasbourg, UNISTRA). “Target-focused library design for efficient hit discovery”.
Prof. Alexandre Bonvin (Computational Structural Biology group, Utrecht University). “"Shape-Restrained Modeling of Protein–Small-Molecule Complexes with HADDOCK".
https://www.youtube.com/watch?v=-6s1Jgx0DSs&ab_channel=ELIXIREurope
June 20th, 17 hs CET:
Dr. Charlotte Minton (Michael Smith Laboratories, University of British Columbia, Canada). “Causes and Consequences of Epistasis in Protein Evolution and Design”
https://www.youtube.com/watch?v=7OCdtCKQS_E&ab_channel=ELIXIREurope
October 14th, 17 CET:
Dr. Joana Pereira, Biozentrum, University of Basel. “Uncovering new families and folds in the natural protein universe”
Dr. Manjeet Kumar, EMBL, Heidelberg. “Eukaryotic Linear Motif (ELM) resource - updates, annotation challenges, and impact of AlphaFold2 and AlphaFold DB on motif-mediated PPIs“. NOT RECORDED
https://www.youtube.com/watch?v=TY2KGnmMorM&ab_channel=ELIXIREurope
21st November, 17 CET:
Dr. Eli Draizen (Postdoctoral Scholar, UC San Francisco). “Is this yet another multiverse talk?! Exploring murky regions of the protein multiverse with ancestral fragments and deep generative models”.
https://www.youtube.com/watch?v=M54oWcJgiVo&ab_channel=ELIXIREurope
Alberto Borsatto ( PhD Student - Gervasio Lab, University of Geneva). “SWISH-X, an expanded approach to detect cryptic pockets in proteins and at protein-protein interfaces”
https://www.youtube.com/watch?v=kQnHB9lOAQ8&ab_channel=ELIXIREurope
- December 5th 2023, 10:0 CET:
Prof. Roland Dunbrack: “Modeling all 437 catalytic typical protein kinases in the human proteome in active form”.
Prof. Michael Stump “The topological properties of the protein universe”.
https://www.youtube.com/watch?v=A71XAiRfDTI&t=39s&ab_channel=ELIXIREurope
The slack community has a total of 377 members as of December 2023.
Equity, Diversification, and Inclusion Efforts
Gender representation: We do not have access to any information relative to the distribution of sex or self-defined gender of participants. We strive to have a balance in terms of sex of keynotes at the level priorities of invitations but in 2022, none of the female (description of sex) invited speakers we able to accept our invitation. One interesting activity that was held during this year’s edition was the Past, Present and Future sessions where attendees were able to discuss the ideas that the chairs had for the future of the COSI and provide feedback about it.
Web site:
https://3dsig-cosi.github.io
Online Community group: https://join.slack.com/t/3dsig-cosi/shared_invite/zt-aasm9l13-etM~ZBDLmVstoCbOQ1f2jw
Word cloud of 3D-SIG topics
Bio-Ontologies Special Interest Group (SIG) covers the latest and most innovative research in the application of ontologies and more generally the organisation, presentation and dissemination of knowledge in biomedicine and the life sciences.
Leadership Structure
COSI representative to committee: Núria Queralt Rosinach
COSI Alternate representative: Maxat Kulmanov
COSI Proceedings Liaison: Tiffany Callahan and Robert Hoehndorf as alternate
COSI Track Chair(s): Tiffany Callahan and Robert Hoehndorf as alternate
COSI Programming Coordinator for ISCBacademy: Núria Queralt Rosinach
Activities Over the Last Year
Bio-Ontologies (#bioontologies; http://www.bio-ontologies.org.uk/) is an ISMB Community of Special Interest (COSI) that covers the latest and most innovative research in the application of ontologies and the organization, presentation, and dissemination of knowledge in biomedicine and the life sciences.
Bio-Ontologies celebrated its 26th year at ISMB 2023, held in-person in Lyon, France on July 23-27, 2023. The Bio-Ontologies is one of the oldest COSIs and provides a vibrant environment for reporting novel methods, exchanging knowledge and sharing experiences on the construction and application of ontologies in the life sciences. The COSI run track offers a constructive environment to nurture discussion of innovative and scientifically sound work that ranges from preliminary to completed, from both young and experienced investigators alike. Bio-Ontologies participants also benefit from a strongly interdisciplinary setting, where ISMB attendees intermingle with members from American Medical Informatics Association (AMIA), the W3C’s Healthcare and Life Sciences Community Group (HCLSCG), and the National COVID Cohort Collaborative (N3C) thereby increasing impact through new collaborations and broader dissemination. Furthermore, the COSI participates in the ISCBAcademy running two webinars every year to both provide to junior and senior researchers another platform for dissemination, instruction and discussion on current topics of interest in academia, and to provide the opportunity to students and junior researchers to chair the session to foster their involvement, empowerment and networking within the community. We aim at broadening the range of talks by inviting industry speakers with current interesting topics for the bio-ontologies community.
Day 1 of the Bio-ontologies track included a keynote delivered by Dr. Ernesto Jimenez-Ruiz on “Ontology Alignment for Life Sciences” where he shared his ideas on the importance of aligning independently developed ontologies. He described several current efforts, including tools, that can be used to resolve this challenge as well as discussed the value of creating more robust benchmarks. This talk drew a sizable in-person and online crowd and elicited a lively discussion. Bio-ontologies also featured many exciting talks. The day featured 6 talks covering updates on existing ontologies, the development of new ontologies as well as resources to make ontologies more approachable for lay audiences. It also included analyses that leveraged ontologies to improve candidate gene prioritization, comparative coexpression, and the identification of novel cell type-phenotype associations. This day also featured a proceedings talk, which presented a novel method for leveraging knowledge graph reasoning to provide explainable predictions of synthetic lethality. The day closed with an awards session and a brief plenary discussion which focused on ways to expand the Bio-Ontologies community and ways to engage existing community members through novel community-based tasks, inviting relevant external speakers for ISCB webinars, and the development of new social media-based avenues for communication.
Day 2 of the Bio-ontologies track included a keynote delivered by Dr. Janna Hastings on “Ontology-based Interpretability for Large Predictive Models” where she discussed the opportunities and challenges around leveraging LLMs for biomedical discovery. This talk also drew a sizable crowd and elicited a lively discussion. Bio-ontologies and BOSC jointly featured 7 exciting talks on tools on new frameworks for building and querying knowledge graphs, frameworks for improving semantic interoperability, and resources to make data modeling more robust.
The COSI sessions also had a social media impact on Twitter and on Slack (69 members currently registered on the bio-ontologies workspace) and our LinkedIn group (Bio-Ontologies; https://www.linkedin.com/groups/13998260/), which currently has 117 members, which represent both academic and industry settings.
Equity, Diversification, and Inclusion Efforts
Our efforts at ISMB:
Gender representation: Bio-Ontology COSI presenters were equally distributed by gender (~50%)
World research representation: The majority of Bio-Ontology COSI researchers were from Europe (43%), Asia (29%), North America (14%), Africa (14%), and South America (0%)
ISMB track summary:
Number of total submissions: 11
Number of proceedings programmed: 1
Number of keynotes speakers: 2
Number of total talks: 6
Number of posters: 9
Sponsorship obtained: NO
Travel fellowships: 0
Web site:
http://www.bio-ontologies.org.uk/
Online Community groups: https://bioontologies.slack.com
https://twitter.com/bioontologies
https://www.linkedin.com/groups/13998260/
https://groups.google.com/g/bioontologies
Word cloud of Bio-Ontologies topics
Bioinfo-core is a worldwide body of people that manage or staff bioinformatics facilities within organizations of all types including academia, academic medical centers, medical schools, government, biotechs and pharmas. Through this wiki and our online discussion lists we discuss many topics that are challenging bioinformatics cores world wide: from IT, new instrumentation, staffing and training bioinformaticians, tools, software, to services for biologists and MD's.
Leadership Structure
Madelaine Gogol, Stowers Institute, United States
Alastair Kerr, CRUK-MI, University of Manchester, United Kingdom
Brent Richter, Partners HealthCare of Massachusetts General and Brigham and Women’s Hospitals, United States
Alberto Riva, University of Florida, United States
Rodrigo Polo, Agriculture and Agri-Food Canada, Canada
Activities Over the Last Year
We focused on the organization of the workshop session at ISMB. We had six speakers from organizations in Brazil, the USA, Canada and the Netherlands. We also recruited our first keynote speaker, Johannes Koster, who provided an overview and updates of the Snakemake workflow management system. This keynote talk was also re-broadcasted at the ISCBacademy event on September 28th and can be found in YouTube: https://www.youtube.com/watch?v=LchfsoH1ITg.
As it is our COSI's tradition, we had workshops following the speaker presentations and the keynote talk. This year, the workshop sessions were very dynamic and included the following topics: project management, core facilities during the pandemic, protected data and knowledge sharing. It was a very rich exchange of ideas were the workshop attendees as we as the organizers learned from the expertise and experience of each other.
Equity, Diversification, and Inclusion Efforts
In the Bioinfo-Core session at ISMB three out of six speakers were women, and four out of six speakers were from underrepresented groups. Besides the USA, we had representation from speakers from Brazil, the Netherlands, Canada and Germany. We hope to work on more inclusive efforts and to attract speakers from countries and groups not traditionally represented in our session and ISMB. We believe there is a lot to learn from the experiences and challenges from colleagues who have the task of establishing core facilities from the ground up in adverse situations and we hope to provide a session at ISMB and a place in our Slack group where they can learn from other colleagues who have core facilities that are a bit more established.
Web site: http://bioinfo-core.org/
Word cloud of BioInfo-Core topics
Data visualization cuts across all areas of computational biology. On the one hand, sophisticated data visualization techniques are required to allow the biologist to explore their large/complex datasets and gain insight from them. On the other hand, this approach can lower the black-box nature of complex (bioinformatics) algorithms. The goal of BioVis is to bring together researchers from the visualization, bioinformatics, and biology communities with the purpose of educating, inspiring, and engaging bioinformatics and biology researchers in state-of-the-art visualization research, as well as visualization researchers in problems in biological data visualization.
Leadership Structure
General Chairs
Helena Jambor, TU Dresden, Germany
Michael Krone, University of Tübingen, Germany
Jan Byška, Masaryk University, Czech Republic
Website Chair
Kari Lavikka, University of Helsinki, Finland
Publicity Chair
Jen Rogers, University of Utah, USA
BioVis COSI @ ISMB 2023 Abstracts Chairs
Katarína Furmanová, Masaryk University, Czech Republic
Aditeya Pandey, Regeneron Genetics Center, United States
Qianwen Wang, Harvard University, USA
ISMB Liaison Chairs
Jim Procter, University of Dundee, UK
BioVis Workshop @ IEEE VIS 2022 Chairs
Barbora Kozlikova, Masaryk University, Czech Republic
Daniel Jönsson, Linköping University, Sweden
Renata Raidou, TU Wien, Austria
Sean O’Donoghue, Garvan Institute of Medical Research, Sydney
Steering Committee
Jan Aerts, KU Leuven and U Hasselt, Belgium
G. Elisabeta Marai, University of Illinois at Chicago, USA
Kay Nieselt, University of Tübingen, Germany
Jim Procter, University of Dundee, UK
Marc Streit, Johannes Kepler University Linz, Austria
Cagatay Turkay, University of Warwick, UK
Honored Members
Michel Westenberg, (passed away April 2021) Eindhoven University of Technology, The Netherlands
Activities Over the Last Year
The BioVis 2023 main event at ISMB hosted two distinguished keynote speakers. Anna Vilanova (TU Eindhoven, NL) examined how accurate analytical models, tailored visual encodings, and carefully designed interactive visualisations allow effective analysis of complex data. Marc Baaden (IBPC Paris, CNRS FR) reviewed the origins of the visual representations, technologies, and software tools used today for molecular graphics and explored its role in teaching, education, science communication, and research. Furthermore, there were 4 long talks, 8 short talks, and 14 posters presented by authors from all over the world. Most notably, Kari Lavikka et al. (University of Helsinki, Finland) received the Best Abstract award for their submission on the development and application of Jellyfish plots in tumour dynamics. Felix Offensperger et al. (Ludwig-Maximilians-Universität München, Germany) received the Best Poster award for their work on visualizing (differential) expression patterns with fuzzy concepts using FlowSets diagrams. Finally, Sean O’Donoghue et al. (CSIRO, U. Sydney, Garvan & UNSW, Australia) received the Test of Time award for their work on the Molecular Control Toolkit, which was originally presented at BioVis 2013.
Another important meeting was the BioVis Workshop @ IEEE VIS 2023. This event serves as a platform for visualization researchers to increase the impact of visualization approaches in biology and medicine. Prior to the workshop, we invited researchers to participate in a visualisation challenge. This year's topic was based on a study recently published in Cancer Cell by the ProCan® team. The tasks were to (1) create an easy-to-understand visualization of peptide and protein intensities by including phenotypic information about tissue and cancer types; (2) design an intuitive visualization that seamlessly connects drug responses to their respective proteins and cell lines or (3) create an interactive tool to visualise drug response data across cell lines in relation to their proteomic data. A total of 3 submissions were evaluated by a jury, and each gave a talk during the workshop. The submission by Brich et al. (University of Tübingen, Germany) was awarded as the winner of the contest. In addition, we invited Drew Berry as the keynote speaker. He presented wehi.tv’s latest productions that generate dynamic, real-time organelle-scale models for science communication and education.
Finally, we continued the online BioVis meetup series, albeit at a reduced frequency, as our focus this year was to fully engage and ramp up the in-person BioVis events rather than virtual seminars. Two meetups have taken place, one discussing visualizations for communicating complex research with the information graphics team of Nature, and one on communicating Scientific Concepts Through Art. We are planning to continue the virtual meetup series with 3-5 seminars per year.
Equity, Diversification, and Inclusion Efforts
BioVis strives to achieve equity and diversity both in its management and operational structures. The BioVis COSI is committed to fostering a diverse and inclusive community. We recognize that diverse perspectives and experiences are critical for advancing science and fostering innovation. We therefore welcome and encourage participation from individuals of all backgrounds and identities, and this was reflected in the diverse participation of scientists hailing from across the world.
We are committed to creating a conference environment that is welcoming, respectful, and inclusive for all attendees. We have ensured that participation and organisation were balanced with respect to junior and senior members of our community and also intentionally gender-balanced the general chairs, keynotes, as well as talks selected from abstracts.
We recognize that equity, diversity, and inclusion are ongoing efforts, and we will work to continuously improve our practices and policies to better support and promote EDI.
Web site:
http://biovis.net/
Online Community groups: https://groups.google.com/a/iscb.org/g/biovis.cosi
biovis2016.slack.com
Word cloud of BioVis topics
Since its inception in 2000, BOSC has provided a forum for developers and users to interact and share research results and ideas in open source bioinformatics and open science. BOSC’s broad spectrum of topics includes practical techniques for solving bioinformatics problems; software development practices; standards and ontologies; approaches that promote open science and sharing of data, reproducible results and software; and ways to grow open source communities while promoting diversity within them. BOSC is run by the Open Bioinformatics Foundation (OBF), a non-profit organization that promotes the open development and dissemination of bioinformatics software and data.
Leadership Structure
The Open Bioinformatics Foundation (OBF), an all-volunteer organization that is responsible for BOSC, is led by an elected Board that currently includes 9 members (https://www.open-bio.org/board/), with Peter Cock as President, Nomi Harris as Vice-President, and Heather Wiencko as Treasurer.
BOSC (https://www.open-bio.org/events/bosc/) continues to be chaired by Nomi Harris, who has chaired or co-chaired the meeting since 2010. For 2024, our organizing committee includes 8 people.
Activities Over the Last Year
OBF accomplishments over the past year include:
- Held public OBF Board meeting in December 2023 (), at which two Board officers and one Member-at-Large were re-elected.
- OBF Event Fellowships: In 2023, 5 people were awarded funds to attend events related to open science (as of December 21, 2023).
- OBF/BOSC presented two ISCBacademy talks:
- March 2023: Hannah Wei, “Re-Thinking the Patient’s Role in a Learning Health System: Lessons from the Patient-Led Research Collaborative” (https://www.open-bio.org/2023/03/07/iscbacademy-webinar-on-patient-led-research/)
- October 2023: Sierra Moxon, “LinkML: an open data modeling framework, grounded with ontologies” (https://www.open-bio.org/2023/09/18/iscbacademy-webinar-oct-3/)
- OBF newsletters went out in February and September 2023: https://github.com/OBF/newsletter/tree/master/newsletters
- OBF and BOSC remained very active on social media. We posted 19 articles on our blog (open-bio.org/blog/) as well as many tweets, emails to our subscribers, and communication in public Slack channels. In 2023, we started OBF and BOSC groups on LinkedIn and Bluesky, and left Twitter/X.
BOSC accomplishments over the past year include:
- BOSC 2023, the 24th annual Bioinformatics Open Source Conference, was part of ISMB/ECCB 2023 in Lyon, France, and online. Approximately 200 people participated in BOSC sessions.
- In addition to 43 talks and 49 posters, BOSC 2024 featured two keynotes: Sara El-Gebali, who spoke about “A New Odyssey: Pioneering the Future of Scientific Progress Through Open Collaboration”, and Joseph Yracheta, who spoke about “The Dissonance between Scientific Altruism & Capitalist Extraction: The Zero Trust and Federated Data Sovereignty Solution.”
- As in 2022, a joint session brought together BOSC and the Bio-Ontologies COSI.
- The conference ended with a panel on Open and Ethical Data Sharing.
- More info about BOSC 2023: https://f1000research.com/articles/12-1568
- BOSC was preceded by CollaborationFest (https://www.open-bio.org/events/bosc-2023/obf-bosc-collaborationfest-2023/), a collaborative work event that brought together about 40 participants to collaborate on 10 different projects, one of which resulted in a peer-reviewed publication: “Making Biomedical Research Software FAIR.”
Equity, Diversification, and Inclusion Efforts
- The OBF Event Fellowship program is specifically aimed at improving diversity at bioinformatics-related events by helping to make events financially accessible to people from groups that are traditionally underrepresented in bioinformatics and open science. Formerly the OBF Travel Fellowship, the program was extended in 2020 to cover costs associated with participating in online events such as virtual conference fees, headsets, high speed internet, etc. (https://www.open-bio.org/event-awards/).
- As in previous years, BOSC prioritized its time and funds to ensure our meeting was more diverse and inclusive. Anyone who submitted an abstracts to BOSC could request fee assistance by checking a box on the submission form. Additionally, participants were encouraged to apply for OBF Event Fellowships. In 2023, 15 people (13 from underrepresented groups) were granted free registration thanks to a combination of registration fee waivers and OBF Event Fellowships.
- This year, we restructured our keynote speaker selection process to make it more inclusive and transparent. In the first phase, we invited as broad a community as possible to nominate potential speakers. Then, we applied our Invited Speaker Rubric to narrow down the list of candidates and solicit community feedback. This process has helped us choose keynote speakers whom we believe represent a diversity of backgrounds and ideas.
- We began offering keynote speaker honoraria last year, in recognition of the fact that not all researchers are privileged to be able to gift their time. Those who did not want to accept honoraria were given the option to donate it back to help support future EDI endeavors. We see the honorarium program as one aspect to our inclusiveness in choosing keynote speakers, which we are still striving to improve. In the past 12 years, BOSC has had 24 keynote speakers. 11 were women or non-binary people; 5 were non-white (all of those in the past 6 years).
- Both of the BOSC 2023 keynote speakers discussed topics related to inclusion and equity. The first BOSC 2023 keynote was delivered by Sara El-Gebali, who spoke inspiringly about “A New Odyssey: Pioneering the Future of Scientific Progress Through Open Collaboration,” with case studies showing how open collaboration can strengthen inclusive scientific communities and vice-versa. The keynote by Joseph Yracheta, an Amerindigenous scientist, discussed our ethical responsibilities as people who work on open source tools and open bioinformatics research to ensure that indigenous data is ethically sourced and used.
- At BOSC 2023, the panel on Open and Ethical Data Sharing centered on the idea that we individually, and our scientific societies, can be advocates for better practices in ethical data sharing.
- A new page on our website (https://www.open-bio.org/bosc-dei/) describes how BOSC has supported diversity, inclusion and accessibility (also known as DEI or EDI).
Web site:
https://www.open-bio.org/events/bosc/
Online Community groups: https://groups.google.com/forum/#!forum/bosc-announce
https://join.slack.com/t/obf-bosc/shared_invite/zt-n5ur1gsj-z2C~69_4lYTFPg5tbWA8Ew
https://genomic.social/@OpenBio
https://genomic.social/@BOSC
https://www.linkedin.com/groups/9539620/
https://www.linkedin.com/groups/14344023/
https://bsky.app/profile/openbio.bsky.social
https://bsky.app/profile/bosc.bsky.social
Word cloud of BOSC topics
In its unique approach through open-ended data analysis contests CAMDA highlights key challenges in analysing massive data in the life sciences.
The increasing relevance of Big Data in the modern life sciences keeps CAMDA topical, addressing one of the grand challenges in the field. Improvements in both the productivity and accessibility of genome-scale assays together with the recognition that high-dimensional profiling requires large sample sizes are driving ever increasing dataset sizes. Still, the data analysis bottleneck limits the rate with which new medical and biological insights can be found. CAMDA tackles this challenge head on, introducing and evaluating new approaches and solutions. Specifically, the conference presents new techniques in bioinformatics, data analysis, and statistics for large data sets, combining multiple data sources and effective computational inference.
Leadership Structure
CAMDA is co-chaired by Wenzhong Xiao (Harvard/Stanford), Joaquin Dopazo (Spanish Health Ministry), David Kreil (Vienna), and Pawel Labaj, a young group leader at a top university in Poland.
The chairs are supported by a Scientific Committee which features a good mix of senior established researchers like Wolfgang Huber and young talent in the field. We are actively working to invite additional young promising members, with Maria Pena Chilet recently joining us, a highly cited young researcher from Andalusia.
Activities Over the Last Year
The challenges for CAMDA 2022 were advertised early in the year and included:
- The Extended Literature AI for Drug Induced Liver Injury Challenge provides biomedical publications curated by FDA experts on DILI. Build their digital twin to identify the sparse positives and demonstrate robustness under distributional shifts!
- The Anti-Microbial Resistance Forensics Challenge features diverse metagenomics profiles from urban and non-urban areas. Track emerging AMR and its relationship with phages!
- The Disease Maps to Modelling COVID-19 Challenge provides highly detailed expert-curated molecular mechanistic maps for COVID-19. Combine them with available omic data to expand the current biological knowledge on COVID-19 mechanism of infection and downstream consequences!
Lynn Schriml (University of Maryland, Baltimore, U.S.A.) started the first day with her keynote on Addressing standardization challenges through integrated approaches in biomedical and genomic data. She has reviewed the development and utilization of standards for characterizing complex disease mechanisms and genomic metadata integration. She also has examined sources of unrecognized bias of collected data and highlighted standardized approaches to address these challenges among genomic and cancer resources. Krista Ryon subsequently provided a focus on metagenomics, sharing highlights from recent efforts of the MetaSUB Consortium in global-scale characterization of urban microbiomes and antimicrobial resistance. Both presentations were highly topical for the metagenomics challenges CAMDA has been running since 2017. Looking into the future, Leonid Chindelevitch (Imperial College London, U.K.) presented an outlook on antimicrobial resistance prediction from whole-genome sequence and metagenomic data, which is now actively being considered for one of the upcoming CAMDA 2023 challenges.
In an ISMB Proceeding presentation fitting well into this area, Tianqi Tang explained how a convolutional neural network could predict phage–host matches based on relatively short contigs. This is a key advancement for metagenomics studies, where instead of wholly assembled genomes typically only contigs are available, which can be as short as a few hundred bases. The second part of the session comprised four short talks on data quality and availability, as well as the challenges of benchmarking in the life sciences. These issues are all central for the reproducibility of science, as also discussed in the day’s keynote. Examples highlighted in the short talks included the incompleteness of public metadata accompanying omics studies, inconsistencies in immunogenetics databases across diverse populations, as well as the limited availability of open T-Cell Receptor Repertoire data. A shared-hub for Universal Benchmarking of Machine Learning-Based Protein Functional Annotation was presented as a possible solution towards greater reproducibility in that field.
The next session started with an overview of past metagenomics challenges at CAMDA, study highlights, and novel research projects evolving from them presented by Paweł Łabaj. Following that Deisy Gysi (Harvard & Northeastern, U.S.A.) outlined the pandemic-time application of AI graph network based methods to drug repurposing. The COVID-19 pandemic highlighted the importance of prioritizing candidates from drugs approved already for other diseases for potential repurposing in the treatment of SARS-CoV-2 infections. Her lab led by Barabasi evaluated existing drug-repurposing methodologies and used a consensus algorithm to increase the accuracy of the predictions. The top-ranked drugs were eventually screened for activity in human cells, identifying six drug candidates that reduced viral infection, four of which were considered promising for further investigation. Notably the presented strategy is data efficient and thus has significance beyond COVID-19, allowing the identification of drug-repurposing candidates for neglected or rare diseases. Mitsuhiro Odaka also reported COVID pandemic related research. Although intercellular attachment between cells are potentially significant in COVID-19, the interactions between the molecules involved have received little attention. The limited understanding of how they are regulated in COVID-19 is thus responsible for our incomplete knowledge of relevant signaling pathways. Speaker demonstrated how data-driven and knowledge-based approaches could be combined to gene network inference for pathway construction, repairing and completing known pathways and improving our current understanding of COVID-19 mechanisms. Joaquin Dopazo closed the day with the 20th anniversary presentation CAMDA Highlights - Making History.
Zhiyong Lu (NCBI, NLM, NIH) kicked off the second day of the conference with his keynote PubMed and Beyond: Biomedical Text Mining for Knowledge Discovery, highlighting the power and breadth of applications of modern text mining methods for finding and extracting knowledge from the biomedical literature. The keynote incidentally served as a great introduction to the subsequent session on the Extended Literature AI for Drug Induced Liver Injury Challenge.
The session featured 4 talks on the topic. Jung Hun Oh (MSK Cancer Center) demonstrated how a combination of word2vec and tf-idf embeddings yielded a better performance in the identification of papers relevant to Drug Induced Liver Injury than either alone. Amongst the models trained on this embedding, the best performance was achieved using a linear support vector machine. Already at CAMDA’21 Malik Yousef (Zefat College, Israel) went beyond the ‘bag of words’ concept by instead introducing a ‘bag of topics’ extracted by LDA, and ranked by Random Forest with Monte-Carlo stratified cross-validation, this year further refining their choice of thresholds. Ihor Stepanov (National Academy of Sciences of Ukraine) joined us remotely to share a comparative analysis of information theory based statistical methods and transformer-based machine learning techniques for scientific literature classification. Finally, Xianghao Zhan (Stanford University) examined the reliability of DILI relevance predictions by conformality. Overall, both an ensemble model and a TF-IDF model reached satisfactory classification results but TF-IDF had much lower computational costs. In fact, its high performance on the seven hold-out datasets gave the TF-IDF model the top score in several of the challenge leaderboards.
Last but not least, Eunyoung Kim won this year’s CAMDA Trophy for her presentation of late breaking research on the interpretable prediction of drug-drug interactions from drug-induced gene expression signatures (data featured in the CAMDA CMap Drug Safety Challenges). First, drug induced gene expression patterns are learned from the structure and properties of drug compounds, allowing inference also for drugs without expression profiling experiments. An attention mechanism learns which genes are important for drug interactions. Translation embedding is used to represent side effects, which are learned from an older version of DrugBank, with predictions validated on a newer version.
The final ISMB Proceeding talk at CAMDA introduced a novel hierarchical framework for the discovery of structured associations from paired high-dimensional datasets, efficiently integrating hierarchical hypothesis testing with FDR corrections to reveal significant linear and non-linear block-wise relationships among continuous or categorical data.
The last session featured the CAMDA Trophy Ceremony with the first prize going to Eunyoung Kim from the Gwangju Institute of Science and Technology in South Korea. Xianghao Zhan from Stanford University and Ihor Stepanov from The National Academy of Sciences of Ukraine tied for a double-second prize right. The competition was tight, and a an honourable mention went to Jung Hun Oh from Memorial Sloan Kettering Cancer Center in USA. The conference then concluded with discussions about new directions and future CAMDA challenges in the traditional CAMDA Caffé.
Throughout the year CAMDA hosted dedicated Google Group based forums for the different CAMDA challenge communities. CAMDA2021 Trophy Winners were invited to present their latest research in the ISCB Academy webinar series. On October 2021 Roumyana Yordanova of Hokkaido University and the Bulgarian Academy of Sciences shared her Alternative approach for discovering relationship between bacteriophages and antimicrobial resistance in the ISCB Academy, and March 2022 Xianghao Zhan from Stanford University showed how to effectively Filter Drug-induced Liver Injury (DILI) Literature with Natural Language Processing and Ensemble Learning.
Equity, Diversification, and Inclusion Efforts
Among keynotes of CAMDA 2022 half were female: Lynn Schriml, a President of the Genomic Standards Consortium and leader at the Human Disease Ontology project, University of Maryland, Baltimore, U.S.A. In addition, we invited: a presentation by Krista Ryon, a young executive director of MetaSUB International Consortium and a researcher at Weill Cornell Medicine, NY, U.S.A,; and presentation by Deisy Morselli Gysi, a young but already successful researcher from Harvard Medical School & Northeastern University, U.S.A.
CAMDA featured speakers from: France, Israel, Japan, Poland, Spain, South Korea, U.K., Ukraine, and U.S.A.
Web site:
https://www.camda.info
Online Community group: https://groups.google.com/a/iscb.org/g/camda.cosi
Word cloud of CAMDA topics
The CompMS group promotes the efficient, high quality analysis of mass spectrometry data through dissemination and training in existing approaches and coordination of new, innovative approaches. The CompMS initiative aims to exploit synergies between different application domains, in particular proteomics and metabolomics.
Our mission is to:
- build a community of scientists working in computational mass spectrometry,
- integrate experimental and theoretical research, and
- bridge the gap between proteomics, metabolomics, and other MS-driven fields.
Leadership Structure
The CompMS COSI is led by a mix of early-career researchers from academia and industry. The current CompMS COSI leadership consists of Wout Bittremieux (University of Antwerp, Belgium), Timo Sachsenberg (University of Tübingen, Germany), Isabell Bludau (Max Planck Institute of Biochemistry, Germany), and Lindsay Pino (Talus Bio, WA, USA).
Activities Over the Last Year
The CompMS COSI significantly contributed to the ISMB/ECCB conference in Lyon, France, through a dedicated track highlighting advances in computational mass spectrometry across proteomics, metabolomics, and lipidomics. The event showcased a dynamic program of talks and posters, headlined by esteemed keynote speakers: Prof. Dr. Laura Goracci delved into next-generation lipidomics with Lipostar 2, Prof. Dr. Bernhard Renard explored machine learning to understand proteomics data, and Prof. Dr. Tomáš Pluskal presented on enhancing mass spectrometry data processing in MZmine 3. The conference featured 9 oral and 23 poster presentations, with Cecile Le Sueur and Yannek Nowatzky receiving awards for their outstanding contributions.
Beyond the conference, CompMS COSI members actively participated in global forums, contributing to major events, including organization of the Bioinformatics Hub at the American Society for Mass Spectrometry and Human Proteome Organization conferences, participating in the Asilomar Conference on Mass Spectrometry and Computational Mass Spectrometry, and imparting knowledge through educational workshops, including Skyline software training and a proteomics course at Cold Spring Harbor Laboratory.
Marking a new chapter in 2023, the CompMS COSI launched a community Slack workspace, rapidly growing to over 250 members. This digital platform fosters vibrant discussions and networking among professionals and enthusiasts in the computational mass spectrometry field. The collective efforts and engagements of the CompMS COSI are pivotal in driving innovation and collaboration within this scientific community.
Equity, Diversification, and Inclusion Efforts
We are committed to diversity, actively seeking balanced gender and geographical representation in all CompMS activities. This commitment was evident at the ISMB/ECCB 2023 conference, where our keynote lineup featured one female and two male speakers, reflecting the 2:1 male-to-female scientist ratio observed in accepted oral and poster presentations. We pledge to continuously strive for gender balance among our keynote speakers, aiming to at least match or exceed the community's gender proportion and pushing towards parity with thoughtful and strategic invitations.
Additionally, we especially emphasize advancing the work and visibility of early-career scientists. In a significant stride towards inclusivity, 88% of accepted talks were delivered by early-career researchers, underlining the COSI's dedication to fostering the next generation of scientific leaders. Further cementing our commitment to diversity and inclusion, we are proud to have recently established a Diversity Grant in computational mass spectrometry, funded by the Chan-Zuckerberg Foundation. This initiative aims to enhance diversity and representation within the community, ensuring a rich tapestry of perspectives and ideas in computational mass spectrometry.
Web site:
http://compms.org/
Online Community group: https://compms.slack.com/
Word cloud of CompMS topics
The ISCB EDUCATION COSI focuses on bioinformatics and computational biology education and training across the life sciences. A major goal of this COSI is to foster a mutually supportive, collaborative community in which bioscientists can share bioinformatics education and training resources and experiences, and facilitate the development of education programs, courses, curricula, etc., and teaching tools and methods.
Leadership Structure
The COSI is managed by three co-chairs: Patricia Palagi, Russell Schwartz, and Wai Keat Yam.
Activities Over the Last Year
The most prominent activity of the COSI each year is running the Education Track of ISMB. For ISMB 2023, this was an exciting program of invited and contributed talks as well as a panel discussion and an education component of the poster program. Last year's program included keynote talks by Eija Korpelainen on evolving training needs and Mohammed Asif Khan on grand challenges for bioinformatics education in the coming decades, a follow-up panel discussion on the grand challenges, a special session on feedback from the Bioinformatics Education Summit (see below), a proceedings talk on education hackathons, and a variety of contributed talks and posters.
The COSI also overlaps with the leadership of the Workshop in Education in Bioinformatics (WEB), a more interactive form of education session run typically just before or after the Education COSI track each year. COSI funds each year help to support inviting keynote speakers to the WEB. This year, that program consisted of two keynote speakers on the topic of accessibility, Alex Lu and Jurgi Giraud, and interactive discussions and breakout sessions on accessibility and inclusivity/equity in training.
The COSI has also been one of the participating organizations each year of the Bioinformatics Education Summit, an annual meeting that has become the premiere event of the international bioinformatics education community. The Summit is a working meeting, currently running in hybrid mode, where sessions are devoted to working groups developing materials and strategies to assist people working in education and training in bioinformatics. COSI members were participants and organizers of several sessions on topics such as developing core competencies and assessing degree programs and short courses.
The COSI also works to encourage education-related events in other ISCB-affiliated conferences. This typically takes the form of some form of special session or panel discussions COSI members will offer as part of the program of other meetings. One such session in the past year consisted of a virtual co-organized session of the Bioinformatics Education Summit at the ISCB-affiliated Great Lakes Bioinformatics (GLBIO) meeting, which happened to occur simultaneously.
The COSI also supports ISCB education efforts by recruiting speakers each year for the ISCB Academy and having members who serve as organizers of the ISMB Tutorial program. The most recent ISCB Academy talk was from Stephen Piccolo, who was selected to expand on a very well received contributed talk of his at the preceding ISMB on the use of large language models in bioinformatics education.
Equity, Diversification, and Inclusion Efforts
EDI efforts are always a priority of the COSI in organizing its major activities, notably in ensuring outreach to diverse groups for recruiting keynote speakers each year. This is reflected in a slate of speakers that has reflected the diversity of the organization by a variety of measures, for example in an exceptionally good gender balance and geographic diversity compared to other society events. It is also reflected in the selection of topics for COSI sessions, for example in emphasis on training in low and middle income countries.
Web site:
https://www.iscb.org/education-and-training-resources
Online Community group: https://groups.google.com/a/iscb.org/forum/#!forum/education.cosi
Word cloud of Education topics
Evolution and comparative genomics are deeply intertwined with computational biology. Computational evolutionary methods, such as phylogenetic inference methods or multiple sequence alignment, are widely used yet remain far from “solved,” making them intense and active areas of research. At the same time, evolutionary and comparative genomics are inherently transdisciplinary in that many biological research domains have some evolutionary component (e.g., cancer genomics, epidemiology, toxicology, population genetics, functional genomics, microbial genomics and ecology, and structural biology). The scope of this COSI remains broad intentionally. The track features a mix of proceedings, highlighted, and invited talks. Priority is given to contributions relevant to multiple areas of applications or topics not covered by other specialized COSIs.
Leadership Structure
• Nadia El Mabrouk, COSI co-Chair (University of Montreal), http://www.iro.umontreal.ca/~mabrouk/
• Janani Ravi, COSI co-chair (University of Colorado Anschutz Medical Campus), https://jravilab.github.io
• Erin Molloy, Abstract co-chair (University of Maryland), https://www.umiacs.umd.edu/people/emolloy
• Edward Braun, Abstract co-chair (University of Florida), https://people.clas.ufl.edu/ebraun/
• Katharine Jahn Abstract co-chair (University of Berlin)
• Wataru Iwasaki Abstract advisor (University of Tokyo)
• Dannie Durand Abstract advisor (Carnegie Mellon University) https://www.cmu.edu/bio/people/faculty/durand.html
• Giltae Song Award co-chair (Pusan National University)
• Louxin Zhang Award co-chair (National University of Singapore) https://discovery.nus.edu.sg/822-louxin-zhang
• Aida Ouangraoua, Communication/website (University of Sherbrooke) https://www.chairs-chaires.gc.ca/chairholders-titulaires/profile-eng.aspx
• Alexey Markin, Webinar (postdoc at USDA), https://markin-alex.github.io/publications.html
Activities Over the Last Year
This year, our Evolution and Comparative Genomics (EvolCompGen) COSI, organized on two full days, reached a number of presentations, posters, and attendance largely exceeding that of previous years. We enjoyed 22 proceedings, short, and long talks each of the two days, for a total of 44, in addition to 52 posters being presented during the poster session. A few presentations were given online, but most of them were presented on-site.
During the workshop, we learnt about the fractionation of somatic cells, the suppression of recombination in sex chromosomes, horizontal gene flow, novel approaches for aligning CRISPR arrays, and more applied and timely studies on human pathogens, cancer evolution, and cultivated beets, to mention only a few. We also heard about the latest algorithmic and software developments for phylogeny estimation, prediction, and evolutionary study of genes, gene syntenies, metabolic networks, as well as transposable elements.
For the significance and novelty of their work, as well as the visual appeal and flow of their presentation and poster, awards were nominated to Brice Letcher and Gwanghoon Jung; with honorable mentions to Xiaoyue Cui, Axel Fehrenbach, and Sarah Farhat; the best poster award went to Daehong Kwon; with honorable mentions to Vivian B. Brandenburg, Sean MacRae, and Xiaoyue Cui.
Throughout the year, the EvolCompGen outreach and research community engagement continues mainly through webinars, but also communications through Twitter: @EvolComp, Web: evolcompgen.org, and Slack: bit.ly/join_evolcompgen. We thank Erin Molloy for her great webinar organization throughout the last years, and welcome Alexey Markin, who’ll take over this year, with a first webinar on November 10th given by Sungsik Kong from the University of Wisconsin-Madison on phylogenetic network estimation from site patterns using composite likelihood.
Equity, Diversification, and Inclusion Efforts
The diversity of our COSI is perfectly reflected in the leadership structure, whose members are early career and established professors from all corners of the world (Asia, Europe, United States), with an equal distribution of men and women. In order to ensure that we have as much diversity in the next meeting, we will make sure to use available COSI funds to help support the travel costs of students who are located at significant distances from the conference venue, in particular, students from Asia and Africa. In addition, it would be great to give registration/travel waivers to students and PIs from low-income and developing countries. Also, more effort into accessibility of all kinds, vision/hearing difficulties would be most helpful to ensure better inclusion.
Web site:
https://evolcompgen.org
Online Community group: https://ismb-evolcompgen-cosi.slack.com
Word cloud of EvolCompGen topics
The accurate annotation of protein function is key to understanding life at the molecular level. The Function SIG COSI brings together computational biologists, experimental biologists and biocurators who are dealing with gene and gene product function inference and annotation, to share ideas and create collaborations. The Function-SIG holds annual meetings and conducts the multi-year Critical Assessment of protein Function Annotation, or CAFA, experiment.
Leadership Structure
Iddo Friedberg, Dukka KC, Ana Rojas, Kim Reynolds, Mark Wass: Function COSI co-organizers
Iddo Friedberg, Predrag Radivojac, Sean D Mooney: CAFA (Critical Assessment of Function Prediction) challenge co-organizers.
Activities Over the Last Year
In late 2022 the success and scope of the CAFA challenges drew the attention of Kaggle, a highly popular data science competition platform and online community of data scientists and machine learning practitioners (https://kaggle.com). The fifth round of CAFA was held in 2023 on the Kaggle platform, with an incentive of $50,000 in total prize money contributed by Kaggle. CAFA 5 was highly successful, with 1,987 competitors organized in 1,625 teams providing 2,850 competing entries. The reason for such a large number of competitors is the broad reach of Kaggle in the machine learning community to advertise this challenge, and of course the prize money donated by Kaggle to the competition.
Walter Reade, a data scientist from Kaggle who worked on CAFA over the past year gave a talk on “Crowdsourcing (Data) Science on Kaggle”. Yana Bromberg from Emory University and Rutgers University was our other keynote speaker and she talked about “Learning from unpopular activities: can unknown functions guide exploration of microbiome environmental preferences?”
Equity, Diversification, and Inclusion Efforts
The Function COSI considers EDI when identifying potential keynote speakers, and when obtaining funding for URM students and trainees that affords EDI considerations.
Web site:
https://www.biofunctionprediction.org/
Online Community group: https://biofunctionprediction.org
Word cloud of Function topics
HiTSeq is devoted to the latest advances in computational techniques for the analysis of high-throughput sequencing data including novel algorithms, analysis methods and applications in biology where high-throughput sequencing data has been transformative. It provides a forum for in depth presentations of novel algorithms, analysis methods, and applications in multiple areas of biology that HTS is transforming.
Leadership Structure
Ana Conesa, PhD
Spanish National Research Council, I2SysBio, Paterna, Spain and University of Florida, Gainesville, FL, USA.
Francisco M. De la Vega, D.Sc.
Stanford University and Tempus AI, USA
Kjong Lehman, PhD
Centre of Medical Technology, Aachen, Germany
Christina Boucher, PhD
University of Florida, Gainesville, FL, USA.
Kristoffer Sahlin, PhD
Stockholm University, Stockholm, Sweden
Brona Brejová, PhD
Comenius University in Bratislava, Slovakia
Can Alkan, PhD
Bilkent University, Ankara, Turkey
Dirk Evers, PhD
Dr. Dirk Evers Consulting, Enger, Germany
STUDENT -PhD Student- Ana Isabel Castillo Orozco MSc
McGill University, Montreal, Canada
Activities Over the Last Year
- New organizing committee members. Brona Brejová and Kristoffer Sahlin were recruited to expand the panel's expertise in various areas of sequence analysis. A student organizing committee member (Ana Isabel Castillo Orozco) was recruited to broaden expertise and outreach among student organizations.
- In preparations for the HitSeq Track (ISMB/ECCB, 2023, Lyon, France), we acquired two keynote speakers.
- ISCB Academy: Ana Conesa gave a talk at the ISCB Academy about “Exploring the Potential and Risks of Third-Generation Long-Read Transcriptome Sequencing for Unraveling Transcriptome Complexity”. The talk focused on the new applications of long-read sequencing, the method of the year 2022, applied to the transcriptome. A 10-minute QNA session followed the talk.
- HitSeq Track (ISMB/ECCB, 2023, Lyon, France):
We designed a program that spanned several key areas in High throughput sequence analysis, including long-read processing, structural variation detection, methods for cancer genomics, indexing and querying of large sequence datasets, metagenomics species classification, pangenome analysis, and spatial transcriptomics analysis.
On the first day of the conference, Irene Papatheodorou delivered a keynote to a fully packed conference room on utilizing single-cell data to comprehend disease and cell type variations across species. The presentation highlighted tools and data resources on the EMBL-EBI expression atlas website, emphasizing the Single Cell Expression Atlas. Irene showcased different views on the data through queries, encouraging the integration of data from various studies. The talk also covered cell-type deconvolution methods across species, including a comprehensive benchmarking of methods. Ongoing projects, such as consensus deconvolution for Gene Expression Atlas, were mentioned. Cross-species integration challenges were discussed, with a focus on tools working well for closely related species. Questions from the audience touched on topics like criteria for RNAseq data inclusion, benchmark dataset availability, cross-species integration challenges, and the evaluation of treatment side effects.
Other presentations focused on the latest algorithms to correct high throughput sequencing technologies, such as high nanopore error rates, and algorithms to detect genomic variation. Alexander Schoenhuth's discussed error correction in long reads using variation graphs with VeChat. Can Firtina presented RawHash for real-time analysis of Nanopore signals. Alister D'Costa's presented SomarvG, an algorithm for the detection of chromosomal translocations. Attendees inquired about the resolution of SomarvG and performance on nested variations.
After lunch, Antonio Collesei introduced ALLSTAR for inferring causal rules between somatic mutations and cancer phenotypes. Arnab Chakrabarti addressed estimating mutational signature exposure from sparse clinical sequencing data, and Yeremia Gunawan Adhisantoso presented PEKORA for high-performance 3D genome reconstruction. Laura Martens discussed the importance of open chromatin in scATAC-seq analysis. Attendees inquired about the difficulty with handling correlation in scATAC-seq analysis. Last out in the session was Derek Aguiar, who highlighted the significance of non-B DNA structures, such as G-quadruplex, in gene regulation and telomere stability. This represents one of the earliest works done in this field. He discussed methods for predicting non-B structures, both experimental and computational, and explored the correlation between non-B structures and sequencing throughput. Using PacBio sequencing, he detected signatures of non-B DNA. The computational challenges in processing translocation times were addressed, revealing that non-B DNA translocation times significantly differ from B DNA translocation times.
After the afternoon break, Timothé Rouzé presented SuperSampler, addressing the scalability challenge in bioinformatics due to the vast data generated by high-throughput sequencing. SuperSampler employs adaptive sketches with super-k-mers, significantly reducing resource usage while ensuring accuracy and robustness. Other presentations covered various state-of-the-art methods for efficient sequencing data representations for large datasets, such as minimal perfect hashing of k-mers, pangenome alignment index construction, the impact of spaced k-mers on alignment-free genotyping, seeding with minimized subsequences, and scalable sequence database search using Partitioned Aggregated Bloom Comb-Trees. These talks sparked engaging discussions on the possibility of generalizing these approaches for various sequencing datasets, strategies for pangenome alignment index construction, and memory and runtime practicality of alternative seeding strategies.
The second day of the conference featured a keynote presentation by Jan Korbel, focusing on deciphering genomic disease mechanisms through single-cell and single-molecule sequencing. Korbel discussed the long-read based discovery of structural variations (SVs) in a thousand genomes, exploring rearrangement mechanisms and their consequences in single cells. He emphasized the accurate typing of cell identity using nucleosome occupancy based on Strand-seq, highlighting its applications in elucidating disease mechanisms and phenotypes.
Several presentations followed, including Coriolis from Andrew Mikalsen, enabling metagenomic classification on lightweight mobile devices, Metabuli (Jaebeom Kim) achieving high specificity and sensitivity in metagenomic classification through joint analysis of amino-acid and DNA sequences, HaploDMF for viral haplotype reconstruction from long reads using deep matrix factorization, SVJedi-graph improving genotyping of close and overlapping structural variants with long reads using a variation graph, Taxor for fast and space-efficient taxonomic classification of long metagenomic reads, and Themisto (Jarno Alanko), a scalable colored k-mer index for sensitive pseudoalignment against hundreds of thousands of bacterial genomes. Taxon and Themiso are examples of some of the very recent advancements in algorithmic sequencing bioinformatics applicable to metagenomics. Questions on scalability and applications to various microbiome analyses were asked, suggesting that the attending audience also had overlapping interests with the microbiome COSI. Around 150 people attended the sessions on the second day.
Other presentations covered multi-locus approaches for accurate variant calling in low-copy repeats by Timofey Prodanov, variational inference for single-cell transcriptome with DNA barcoding reconstructing unobserved cell states and differentiation trajectories by Koichiro Majima, GAN-based data augmentation for transcriptomics, visualizing spatial transcriptomics with U-CIE color encoding. Corentin Thuilliez, presented CellFromSpace, a versatile tool for spatial transcriptomic data analysis through reference-free deconvolution and guided cell type/activity annotation.
The day concluded with a presentation on demuxSNP for supervised demultiplexing of single-cell RNA sequencing data using cell hashing and SNPs, highlighting the benefits of multiplexing for sequencing more cells at reduced cost and addressing questions about the hashing quality and its variation between biological samples. The conference also featured discussions on leveraging evolutionary constraints to refine somatic variant calls from single-cell sequencing data, emphasizing the challenges of ultra-low depth single-cell whole-genome sequencing data and the need for manual variant call refinement to address scalability issues and accurately capture subclonal variants.
Equity, Diversification, and Inclusion Efforts
To support DEI efforts, we aimed to promote a rough balance between male and female speakers at the annual HiTSeq meeting. For example, we had one female and one male keynote speaker. As for the remaining 28 talks, 8 of the presentations had female first authors, although only 6 of them attended to present. As for the poster submissions, there were 32 (out of 108) submissions with a female first author, where 4 of the 32 submissions were accepted as talks. Some further work remains to achieve a better gender balance.
Web site: http://hitseq.org
Word cloud of HiTSeq topics
The iRNA COSI aims to bring together computational and experimental experts that drive new advances in RNA biology and overcome the computational challenges involved. The iRNA meeting at ISMB focuses on two major areas: (1) the development of computational and high-throughput experimental methods, and (2) the application of such methods to break new grounds in the study of RNA biology and disease. Through the meetings and various online activities we aim to educate and inspire researchers in the field, novice and seasoned alike, by meshing together different aspects of Computational RNA Biology, and promoting cross-disciplinary collaborative research.
Leadership Structure
Organizing Committee
Athma Pai, UMass Medical School, USA
Michelle Scott, Université de Sherbrooke, Canada
Hagen Tilgner, Weill Cornell Medicine
Abstract Committee
Athma Pai, UMass Medical School, USA
Michelle Scott, Université de Sherbrooke, Canada
Hagen Tilgner, Weill Cornell Medicine
Maayan Salton, Hebrew University, Israel
Proceedings Liaison
Jérôme Waldispühl, McGill University, Canada
Activities Over the Last Year
Highlights of the RNA COSI at ISMB 2023:
The 2023 iRNA COSI took place on July 26 and 27. Our first iRNA track keynote talk was given by Irmtraud Meyer from the Berlin Institute for Medical Systems Biology who discussed several computational methods to study the dynamic RNA structure, forming as the RNA is being transcribed rather than assuming a fixed molecule. This talk was followed by several talks discussing optimized RNA design for vaccines, RNA interactions, RNA motifs, small RNAs and transcription start sites. Our second keynote, Julien Gagneur from the Technical University of Munich described the detection and prediction of aberrant splicing using machine learning algorithms before switching to discussing codon usage, relation to energy consumption and its effect on RNA stability. We ended the day with concise and entertaining flash talks advertising some of our posters and a lively poster session followed by our COSI dinner at the Park Brasserie. Our traditional dinner quiz was a great success, probing the depth of our knowledge of local trivia. Our second day featured talks on RNA localization, kinetics, CRISPR design, transcriptome reconstruction, RNA modification and RNA maturation including splicing and polyadenylation and their deregulation in disease. Our first keynote of the day, Maayan Salton from the Hebrew University discussed often disregarded but important regulatory relationships between promoters/enhancers and alternative splicing. The final keynote talk presented by Hagen Tilgner from Weill Cornell discussed recent advances in long reads technology and findings in spatial splicing analysis focused on mouse and human brains. Overall, the meeting provided an overview of new tools, technologies and many fascinating studies leading to a deeper exploration and increased understanding of the transcriptome.
Year-round activities
-Management of the iRNA COSI website to inform and mobilize the community: https://irbgroup.org/
-iRNA talks in the ISCB Academy: Christina Akirtava (CMU) discussed False-positive IRESes, mRNA annotation errors and a paradigm 'unshift' in mammalian development in November 2022 and Dr Matvei Khoroshkin (UCSF) talked about A systematic search for RNA structural switches across the human transcriptome in May 2023.
-Promotion of other meetings involving computational RNA research: EMBL-EBI course on Introduction to RNA-seq and functional interpretation and RiboClub 2023.
Equity, Diversification, and Inclusion Efforts
When choosing our COSI keynote speakers, in addition to well-known scientists in our field, we spend time searching for members of our community whose work we could highlight, who conduct excellent research but might not be senior sought-after speakers (yet), a good proportion of which are members of minorities. When selecting abstracts for our talks, we always consider simultaneously EDI and scientific excellence. Including all researchers that see themselves as computational RNA scientists regardless of their origin is a fundamental principle of our COSI and we are completely convinced that increasing diversity in members increases diversity in ideas, enriching our community.
Web site: https://irnacosi.org/
Word cloud of iRNA topics
Transitioning from a post-doc to a junior PI can be a challenging process requiring careful planning. Once running a group, junior PIs are faced with many new tasks, some of which are learnt on the job. The Junior Principal Investigators group (JPI) aims to provide support during this process via a community of peers.
Leadership Structure
Yesid Cuesta Astroz , Claudia Gonzaga Jauregui, and Melike Donertas serve as the JPI co-leads.
Activities Over the Last Year
The JPI COSI maintains a New PI Slack ISCB channel and participates in the ISCB webinar series. Typically, the JPI COSI strives to plan and moderate a track session and a Birds of a Feather session at ISMB annually. We have now recruited co-leads based in North America and Europe and recommend this moving forward to better facilitate sustained international community-building efforts.
Equity, Diversification, and Inclusion Efforts
As part of ISCB, the JPI COSI values diverse and inclusive participation across the computational biology and bioinformatics community. To that end, our goal is to support community building that is broadly accessible by facilitating virtual and in-person networking, education, and scientific events.
Web site:
http://cosi.iscb.org/wiki/JPI:Home
Online Community group: https://forms.gle/Zaqy1XTM9fiG2Jb59
The MICROBIOME Community of Special Interest aims at the advancement and evaluation of computational methods in microbiome research, especially metaomic approaches. Based on the Critical Assessment of Metagenome Interpretation (CAMI), the COSI supplies users and developers with exhaustive quantitative data about the performance of methods in relevant scenarios. It therefore guides users in the selection and application of methods and in their proper interpretation. Furthermore, the COSI provides a platform for exchange and networking between method developers, and provides valuable information allowing them to identify promising directions for their future work.
Leadership Structure
Alice McHardy; Helmholtz Centre for Infection Research; Germany
Alexander Sczyrba; University of Bielefeld; Germany
Activities Over the Last Year
The MICROBIOME COSI 2023 featured two keynote talks given by (1) Karoline Faust from the KU Leuven, Belgium, on "From hairballs to hypotheses: network analysis applied to microbiome data", and (2) Nicola Segata from the University of Trento, Italy, on "Computational metagenomics to mine the hidden diversity of the human microbiome". On May 16, 2023, we hosted the ISCB academy webinar "Precise Microbiome Genomics to Precise Medicine", by Ami Bhatt from Stanford University, USA.
Equity, Diversification, and Inclusion Efforts
The COSI invited an equal number of male and female keynote speakers.
Web site:
https://www.microbiome-cosi.org/
Online Community groups: https://groups.google.com/g/microbiome-cosi
https://twitter.com/microbiome_cosi
Word cloud of MICROBIOME topics
MLCSB is a community for researchers interested in the interface of data sciences and life sciences, in particular the method development and application challenges that arise for Machine Learning in Computational and Systems Biology.
Leadership Structure
Magnus Rattray and Chloe-Agathe Azencott are MLCSB co-presidents and are supported by the MLCSB board. New co-presidents will be appointed this year (from March).
Activities Over the Last Year
This year's MLCSB track included a panel session on Biological Foundation Models, chaired by Anshul Kundaje who was joined by Bo Wang, Mo Lotfollahi and Dana Pe'er. As well as running the MLCSB COSI, we also support the MLCB meeting which is now a standalone meeting, typically US-based, with several MLCB organisers on our board. We run a COSI twitter account, managed by Sayane Shome, to advertise relevant events and activities.
Equity, Diversification, and Inclusion Efforts
MLCSB encourages diversity and representation within its leadership. The MLCSB meeting has had at least one female co-chair (of two) since 2019 and has female and male co-presidents. Chairs are invited to become board members after ISMB and this ensures the board also has good gender representation. We also pursue increasing diversity in terms of ethnicity and geography in our leadership by ensuring that our track chair selection lists are diverse, which then feeds into our board.
Web site:
http://cosi.iscb.org/wiki/MLCSB:Home
Online Community group: https://groups.google.com/a/iscb.org/g/mlcsb.cosi
Word cloud of MLCSB topics
As more research fields turn to network visualization and analysis for perspective, our Network Biology Community serves to introduce novel methods and tools, identify best practices, and highlight the latest research in the growing and interconnected field of network biology.
Leadership Structure
Committee Representative: Carlo Vittorio Cannistraci, Tsinghua University, China
Alt. Committee Representative: Anaïs Baudot, Aix Marseille University, France
Track Chair: Tijana Milenkovic, University of Notre Dame, US
Track Chair: Marinka Zitnik, Harvard University, US
Proceedings Liaison: Tijana Milenkovic, University of Notre Dame, US
Abstract Chair: Martina Summer Kutmon, Maastricht University, NL
Abstract Selection Committee: Scooter Morris, UCSF, US
ISCBacademy program coordinator: Alexandros Xenos, Barcelona Supercomputing Center, Spain
Liaison ISBC Student Council: Pradeep Eranti, Université de Paris, France
Website Manager: Noel Malod Dognin, Barcelona Supercomputing Center, Spain
Activities Over the Last Year
The essential activity in 2023 was organizing the full-day NetBio COSI session at ISMB/ECCB 2023 in Lyon, France. Highlights included keynotes by Prof. Natasa Przulj and Prof. Desmond Higham, along with six proceedings presentations and seven contributed talks, all in a session that was very well attended, showcasing the interest and enthusiasm for the network biology field. With a record 91 submissions, we also hosted a vibrant poster happy hour showcasing 73 NetBio posters. Congratulations to Océane Cassan (best talk prize) and Salvo Danilo Lombardo (best poster prize).
We are excited that for ISMB 2024, ISCB generously agreed to extend our NetBio COSI session to one and a half days, allowing for an even more comprehensive and engaging program.
We also hosted an inspiring ISCB academy webinar by Michelle Li from Harvard University on 13 December 2023 on “Contextualizing protein representations using deep learning on protein networks and single-cell data”.
Equity, Diversification, and Inclusion Efforts
Careful attention is paid to equity, diversity, and inclusion when selecting committee members and speakers for all organized events and activities. In 2023, a large portion (4/9) of the organizing committee were women. At our ISMB session, we achieved gender balance among our keynote speakers, with one male and one female speaker. Additionally, nearly half (47%) of our other talk and poster presenters were women, showing a significant increase from approximately 35% in 2022 and 29% in 2021.
Web site:
http://cosi.iscb.org/wiki/NetBio:Home
Online Community groups: netbiocosi.slack.com
https://www.linkedin.com/groups/5123610/
Word cloud of NetBio topics
The Regulatory and Systems Genomics Community of Special Interest focuses on computational methods that are important in the study of regulation of genes and systems. The RegSys COSI organizes the following activities: (1) ISMB Regulatory Genomics SIG Meeting and (2) RECOMB/ISCB Conference on Regulatory and Systems Genomics and DREAM Challenges.
Leadership Structure
RegSys COSI Steering Committee
Stein Aerts
Jason Ernst (Co-Chair)
Manolis Kellis
Christina Leslie
Jian Ma
Saurabh Sinha (Co-Chair)
Lonnie Welch (Co-Chair)
Judith Zaugg
Julia Zeitlinger
Activities Over the Last Year
The Regulatory and Systems Genomics Community organizes the following activities each year:
1. ISMB Regulatory & Systems Genomics 2-day Track at ISMB
2. RECOMB/ISCB Conference on Regulatory and Systems Genomics and DREAM Challenges
3. ISCBAcademy - RegSys
4. Online community discussions via Twitter and Slack
Activity summaries are provided below.
1. ISMB Regulatory & Systems Genomics 2-day Track at ISMB
Day 1 (July 25th): The RegSys track opened with Julien Gagneur's keynote presentation on in-silico representations of the regulatory code, focusing on human transcriptional regulatory sequences and self-supervised modelling of genomic sequences. Caroline Uhler delivered another keynote talk on causal representation learning in gene regulation. Other presentations covered diverse topics, including the impact of transcription initiation at microsatellites on gene expression, the differential abilities of Forkhead transcription factors in chromatin engagement, a novel regulatory DNA classifier, and an intrinsically interpretable neural network for genomic sequence-to-function predictions.
Day 2 (July 26th): Mikhail Spivakov led the second day with a keynote presentation probing the relationship between enhancer activity, connectivity, and gene expression. The day also featured Ana Conesa discussing third-generation sequencing technologies for investigating transcriptome complexity. Additional talks included methods for integrating Hi-C and transcriptomics, predicting chromatin interactions, identifying genetic variants affecting genome organisation, multi-modal learning for single-cell multi-omics data integration, and understanding transcriptional repression in safeguarding cell identity.
Overall Summary: The RegSys track at ISMB/ECCB 2023 highlighted key advancements in understanding gene regulation and genome organisation.
The RegSys COSI congratulates the winners of our best talk award: Ibrahim Ihsan Taskiran, VIB-KU Leuven Center for Brain & Disease Research, Leuven, Belgium, for his talk titled “Cell type-directed design of synthetic enhancers,” and Jacob Schreiber, Stanford University, United States, for his talk titled “DragoNNFruit: Learning cis- and trans-regulatory factors of chromatin accessibility profiles at single base and single-cell resolution.”
2. RECOMB/ISCB Conference on Regulatory and Systems Genomics and DREAM Challenges
The 2023 RECOMB/ISCB Conference on Regulatory and Systems Genomics (RSG) was a two-day event that occurred on Nov 28th-29th, Los Angeles, CA, USA and co-located with the meeting on DREAM challenges. RSG was co-chaired by Jason Ernst (UCLA), Ferhat Ay (La Jolla Institute of Immunology/UCSD) and Sushmita Roy (UW Madison). Both days focused on open and emerging problems in gene regulation and regulatory genomics. The conference had a total of five keynotes who discussed cutting edge problems at the forefront of understanding regulatory and systems genomics. The conference also featured contributed talks selected based on abstracts reviewed by a program committee.
Day 1 started in the morning with a keynote from Dr. David Kelley, which talked about the power of deep learning models for interpreting human genetic variation in silico. After lunch, Dr. Jingyi Jessica Li’s talk occurred which focused on the importance of simulated data and accurate data simulators for single cell and spatial transcriptomics and multi-omics. Finally, day 1 ended with a keynote from Dr. David Van Valen on the recent advances of deep learning and foundation models on image analysis and segmentation. Day 1 featured contributed talks in the areas of epigenomics and single cell genomics. Day 2 started with a keynote from Dr. Jean Fan on spatial transcriptomics, followed by contributed talks on genetic variation, disease and network dynamics. Day 2 also featured a keynote from Dr. Remo Rohs who talked about how transcription factors find their target genes genome-wide and the role of DNA shape, followed by contributed talks on the transcriptome and disease.
The conference sessions were moderated by the program chairs a well as other faculty. Poster sessions were organized on both days overlapping lunch allowing formal and informal interactions among presenters and attendees. Overall, RSG gave platform to 5 invited talks, 25 invited talks, 5 poster flash talks and 80 posters over two days of exciting scientific exchange on the beautiful UCLA Campus.
The DREAM session of the conference, composed of 1 full day of sessions, kicked off with the presentation of the results of the The COugh Diagnostic Algorithm for Tuberculosis
CODA TB Challenge, using voice recording to diagnose tuberculosis, using patient data collected in over 5 different countries. It was then followed by a single-cell RNAseq benchmarking challenge, where the depth and number of cells was varied and participants were asked to impute the missing values. The openchallenges.io platform was introduced as a way to centralize the information regarding all the crowdsourcing competitions in Life Sciences. For the DREAM portion, the amazing Keynote by Paul Spellman from UCLA showed how a succession of cancer-specific set of mutations lead to cell de-programing and tumor growth. Results were then presented for The FINRISK HF Microbiome DREAM challenge: Predicting Incident Heart Failure from the Microbiome, a challenge that was technically difficult because this was a model to data scheme, where all models had to be containerized and trained on in silico data. This was followed by the presentations (4 talks) on the Rare-X Challenge which was focused on rare pediatric neurodevelopmental diseases, brought together academic and industry researchers and data scientists to use patient-reported data to address unanswered research questions about rare diseases.
Once again, the enthusiasm of the assistants and organizers shows how important it is for the DREAM community to get together once a year. We were happy to see that the 15th annual DREAM-RSG conference in Los Angeles was a success! The thriving community of UCLA and the location attracted people from all over the world; with attendees all the way from France and South Korea, as well as all over the US. It was fantastic to be able to come together again to discuss science.
3. ISCBAcademy - RegSys
We featured the following DREAM Challenge talks:
A. Brain Tumor Segmentation (BraTS) Continuous Challenge
B. Decision Tree-based models for predicting the efficacy of anti-PD1 treatment
Equity, Diversification, and Inclusion Efforts
Equity, diversity, and inclusion are given careful consideration during the selection of chairs and speakers for all events and activities organized by RegSys COSI. For instance, we have achieved gender diversity in the composition of leadership, committees, and speakers.
Web site:
http://cosi.iscb.org/wiki/RegSIG:Home
Online Community groups: https://twitter.com/ISCB_RegSys
https://regsyscommunity.slack.com
Word cloud of RegSys topics
The Computational Modeling of Biological Systems (SysMod) Community of Special Interest (COSI) of the International Society for Computational Biology (ISCB) is a forum for discussion about the combined use of systems biology modeling and bioinformatics to understand biology and disease. SysMod encompasses all methods used in bioinformatics and systems biology, as well as all biological systems and all applications areas. The main activities of SysMod include an annual meeting at the Intelligent Systems for Molecular Biology (ISMB) conference organized by the ISCB and an online forum.
Leadership Structure
The SysMod COSI is jointly organized by a team of six coordinators, who virtually meet once per month. The team of all current coordinators in alphabetical order includes Shaimaa Bakr (Stanford University, US), Matteo Barberis (University of Surrey, UK), Chiara Damiani (Università degli Studi di Milano-Bicocca, IT), Reihaneh Mostolizadeh (Eberhard Karl University of Tübingen, DE), Bhanwar Lal Puniya (University of Nebraska–Lincoln, US), and Meghna Verma (AstraZeneca, US).
From 2022, Bhanwar Lal Puniya acted as the chairperson and as the primary COSI representative to ISCB. Andreas Dräger acted as the alternate representative to the committee. Meghna Verma is the COSI Programming Coordinator for ISCBacademy. The track chairs for the 2023 annual meeting were Shaimaa Bakr, Matteo Barberis, Chiara Damiani, Reihaneh Mostolizadeh, Bhanwar Lal Puniya, and Meghna Verma. Matteo Barberis and Chiara Damiani were moderators for the 2023 annual meeting.
Activities Over the Last Year
The primary activity of SysMod is the annual one-day meeting at the ISMB/ECCB conference. The event in 2023 featured two keynote talks, 13 contributed talks and 68 posters. In 2023, SysMod organized two ISCBAcademy webinars. Over the past year, SysMod has also maintained a community discussion list (e.g., the announcement of impactful new research, meetings, job opportunities, etc.) and a community calendar (e.g., conferences, workshops, and schools).
Equity, Diversification, and Inclusion Efforts
Over the years, the SysMod coordinators have always promoted equal opportunities for researchers irrespective of their gender, beliefs, ethnicity, and other non-scientific aspects, as the highly diverse composition of the coordinators’ group, the scientific advisory board, and alumni indicate. From its inception, scientific quality has always been the first criterion for inviting keynote speakers. SysMod has always strived to achieve the highest scientific quality while inviting speakers from diverse countries with various backgrounds and ethnicities and with a balance in genders to present a wide variety of excellent researchers within the community. SysMod will continue to support underrepresented researchers when inviting keynote speakers. The submitted contributions, genders, and ethnicities will be balanced, depending on the potential presenting invited authors. Contributed talks and posters, on the other hand, are solely chosen based on scientific quality and soundness.
Web site:
https://sysmod.info
Online Community group: https://groups.google.com/forum/?hl=en#!forum/sysmod
Word cloud of SysMod topics
Text Mining COSI is a community that brings together researchers, computational biologists, bioinformaticians and biomedical text mining specialists seeking to use text mining to advance the biomedical and biological sciences, to bridge the gap between the capabilities of text mining systems and the needs of concrete applications, and to develop new text mining methodologies that advance the state of the art.
Leadership Structure
Representative: Robert Leaman
Track Chairs: Robert Leaman, Lars Juhl Jensen, Cecilia Arighi, Zhiyong Lu
Activities Over the Last Year
The ISMB/ECCB 2023 text mining track was a one-day event featuring two keynote speakers and extensive range of scientific presentations and posters. It also hosted highly attended panel discussion on the implications of large language models, such as ChatGPT, in biology and medicine.
- The COSI track explored methodological advances and developing resources, especially for advanced language models. Recurring themes included extracting relations from the literature, manual annotation for training and evaluation, leveraging weak supervision, and integrating text data with other modalities. Practical applications spanned clinical and bioinformatics research.
- COSI track keynote speaker Martin Krallinger discussed creating natural language processing (NLP) components and resources for shared tasks in collaboration with clinical domain experts. His presentation discussed a spectrum of applications including biomaterial research, cardiology, occupational health and phenotypes, with increasing emphasis on multilingual applications. COSI track keynote speaker Aurélie Névéol discussed the complexities of reproducibility in biomedical NLP, examining how these challenges interact with issues like data privacy. She offered recommendations for enhancing reproducibility for biomedical NLP systems and how NLP systems themselves might help improve reproducibility across science.
- The COSI track ended with a very well attended panel on the applications of ChatGPT and large language models (LLMs) in biology and medicine. Panelist Larry Hunter discussed the tension between the utility of LLMs and how their fluency can be misleading. Panelist Harry Caufield discussed using LLMs to create knowledge graphs, particularly noting recent work mitigating “hallucination” by grounding responses in established ontologies. Panelist Chris Dallego discussed industrial LLM applications in healthcare and science. Audience questions focused on trustworthiness, data protections, addressing bias, democratizing the training and use of LLMs, alongside benchmarking and evaluation.
In January 2023, the COSI hosted Søren Brunak, professor of Disease Systems Biology at the University of Copenhagen, for a ISCB Academy lecture. He shared strategies for text mining to construct disease trajectories at scale by incorporating deep temporal information from patient records.
Equity, Diversification, and Inclusion Efforts
We support diversity in science and seek to create an inclusive community. We consider gender balance and underrepresented backgrounds when inviting keynote and invited speakers, and our scientific program reflects our commitment to promoting researchers from diverse backgrounds.
Web site:
http://cosi.iscb.org/wiki/TextMining:Home
Online Community group: https://groups.google.com/g/text-mining-cosi
Word cloud of TextMining topics
Knowledge-based translational medicine is a rapidly growing discipline in biomedical research and aims to expedite the discovery of new diagnostic tools and treatments by using a multi-disciplinary, highly collaborative, "bench-to-bedside" approach. It involves the integration of multiple high dimensional datasets that capture the molecular profiles of patients, as well as detailed clinical information. Within public health, translational medicine is focused on ensuring that proven strategies for disease treatment and prevention are actually implemented within the community, and on progressing towards data-educated personalised therapy. To genuinely realise the promise of Big Data in healthcare, we must consistently collate the data, annotate it with consistent and useful ontologies, apply sophisticated statistical analysis and translate these findings to the clinic. As a community, we will explore the current status of computational biology approaches within the field of clinical and translational medicine. In this COSI we will bring scientists from both academia and industry to exchange knowledge and foster networking, to help in building up of the translational medicine community.
Leadership Structure
Venkata Satagopam is currently the TransMed COSI Representative and a COSI Track Chair. Maria Secrier is currently the alternate COSI Committee Representative, and a COSI Track Chair. Irene Ong is the COSI Proceedings Liaison and a COSI Track Chair.
Irina Balaur is COSI abstracts chair and a COSI Track Chair, and Heba Sailem, Mansoor Saqi, Bosdriesz Evert serve as COSI Track Chairs. Irina Balaur also serves as the ISCBacademy TransMed Program Coordinator.
Activities Over the Last Year
TransMed 2023 meeting was held during the ISMB/ECCB 2023 conference on July 10 - 14, 2023 in Lyon, France. TransMed 2023 had 2 keynote speakers invited:
Dr Maggie Cheang / Institute of Cancer Research who is the Group Leader of the Integrative Genomics Analysis in clinical trials, Division of Clinical Studies. Her primary research focus is to identify biomarkers that would be clinically relevant. Combining biological knowledge and advanced statistical analytics/machine learning to model the multi-scale multi- “omics data” with clinical outcome, her team has been developing multi-parametric molecular classifiers to predict sensitivity and resistance of tumour biological subtype to therapeutic agents and testing the performance of one of these integrated omics/mathematical algorithms within a Phase III clinical trial at this moment..
Prof Alfonso Valencia, who is ICREA research Professor, Director of the Life Sciences Department of the Barcelona Supercomputing Center, Director of the Spanish National Bioinformatics Institute INB/ELIXIR-ES and coordinator of the data pillar of the Spanish Personalised Medicine intiative, IMPaCT. His research interest is the development of Computational Biology methods and their application to biomedical problems. Some of the computational methods he developed are considered pioneering work in areas such as biological text mining, protein coevolution, disease networks and more recently modelling cellular systems (digital twins).
In total, TransMed 2023 had 1 proceedings, 5 long talks (20 min slot) ,12 short talks (10 min slot) selected from the abstracts and 101 posters.
We had a very exciting day of talks spanning very diverse areas of biology and biomedicine. The day started with. Dr Maggie Cheang’s keynote, who gave a very insightful account of how biomarkers are utilised in the clinic, and the constraints existing around developing and testing biomarkers alongside clinical trials. This was followed by a series of excellent talks covering newly developed statistical, AI and graph reasoning methods as well as applications of existing computational methods to very diverse biomedical datasets, from genomics, transcriptomics, methylation, spatial biology, histopathology, radiology and other clinical markers, to drug data, electronic health records and semantically integrated knowledgebases. We have learned about new methodology to deal with privacy concerns when using genetic data, infer ancestry, understand the effects of genetic alterations or risk factors such as smoking in disease, and also about innovative approaches that we can apply to sparse datasets like those encountered in rare diseases. We also learnt about new ways to identify biomarkers of disease and predict drug effects with implications for understanding clinical failure in drug administration and even in organ transplants. Finally, Prof Alfonso Valencia rounded up the TransMed COSI in the final keynote of the day with a brilliant talk describing how to leverage patient and real world datasets to overcome small sample sizes when studying rare diseases and to maximise the use of recorded data to understand transmission of infectious diseases like Covid. As highlighted by all of today’s speakers, we face increasing analytical and computing challenges when integrating increasingly complex datasets and translating them in the clinic, and there is a clear opportunity for computational, statistical and AI methods to transform the field of translational medicine in the coming years.
We also selected the Best talk and the Best Poster awards considering both the scores of the abstract reviews and presentations during the TransMed 2023 meeting as follows.
Best Talk
Michelle Li, (Harvard Medical School, USA), “Deep learning for diagnosing patients with rare genetic diseases”
Best Poster
Dries Heylen, (UHasselt &VITO NV, Belgium), “Evaluating the frequency of multi-omics longitudinal data collection in the perspective of chronic disease development”
We welcomed new members of the TransMed COSI organizers (core team) for the year 2023-2024:
Evert Bosdriez, PhD. (VU Amsterdam, Netherlands)
We thank very much for her contribution during the 2021 - 2023 time period and said goodbye to
Sanne Abeln, Prof. (VU Amsterdam & Utrecht University, Netherlands)
Equity, Diversification, and Inclusion Efforts
We gave importance to gender balance and had a female and a male keynote speakers forTransMed 2023 COSI. In our COSI organising committee, we have 60% females.
Web site:
https://transmed.github.io/
Online Community groups: https://twitter.com/TransMedISMB
https://www.linkedin.com/groups/8478286/
https://ismbtransmedcosi.slack.com/
Word cloud of TransMed topics
The Variant Interpretation Community of Special Interest (VarI-COSI) is a community of scientists interested in “breaking” the genomic code. The main goal of our COSI is to promote the formation of a collaborative network of scientists interested in the understanding of the meaning of genomic variation as applied to a range of questions, including population studies, functional and evolutionary impacts, and disease.
Leadership Structure
Our COSI is led by 4 co-chairs: Julien Gagneur, Emidio Capriotti, Hannah Carter and Antonio Rausell
Activities Over the Last Year
In the past year, VarI COSI hosted a track on the latest methods for variant interpretation at ISMB/ECCB 2023. The session featured three distinguished keynote speakers, including a local speaker from the WHO Lyons, Matthieu Foll, as well as Jörg Menche (University of Vienna) and Nicky Whiffin (Oxford University), and included 13 talks and 36 posters. The topics addressed during the session encompassed various methodologies (notably with deep learning and network modeling), variant mechanisms (missense and regulatory), and applications (rare diseases, common diseases, and cancer) of genomic variant interpretation. Outside of ISMB/ECCB coordinated a special issue in the International Journal of Molecular Sciences and hosted the June ISCB Academy webinar.
Equity, Diversification, and Inclusion Efforts
We strive to be inclusive of all members of our scientific community and ensure representation at our annual meeting. We consider EDI when identifying potential keynote speakers. The VarI COSI strongly promotes gender balance. Currently Hannah Carter is our representative to the committee.
Web site:
http://varicosi.biofold.org/
Online Community groups: https://groups.google.com/a/iscb.org/g/vari.cosi
https://twitter.com/varicosi
https://varicosi.slack.com/
https://varicosi.biofold.org/
Word cloud of VarI topics