Leading Professional Society for Computational Biology and Bioinformatics
Connecting, Training, Empowering, Worldwide

ISCB News and Announcements

 
 
 

Wednesday, July 23: ISMB/ECCB Day 4 Highlights and Recap
 
 

On day 4 or ISMB/ECCB 2025, we had a full day with 9 COSI tracks, CollaborationFest, and the industry and ELIXIR/NIH-ODSS special sessions. 

Before the ACC was full of attendees, however, another morning started off with the Serene Stretch Symposium. 
 
 
 
 

The ISCB Town Hall was held, allowing attendees to learn about the latest programs and initiatives from ISCB, as well as upcoming conferences. This meeting provided the opportunity for members to provide feedback and suggestions to help shape the future of the society!

Another great highlight of the day was the presentation of the 2025 Outstanding Contribution to ISCB Award to this year's winner, Lucia Peixoto. Lucia was recognized for her leadership, service,
and dedication to building an inclusive scientific community. 
 
 
 
 


**Please make sure to read the “Important Notes” section below.**
 
 
 
 
 

Keynote Address: Charlotte Deane
 
 
 
 

In her keynote address, “Building the future of AI-driven structure-based drug discovery,” Professor Charlotte Deane explored how artificial intelligence is reshaping the development of therapeutics while offering a candid look at both its potential and its pitfalls.

Drug development remains an extraordinarily costly and time-consuming process, with high failure rates and limited incentives to pursue treatments for diseases that disproportionately affect underserved populations. Deane made a compelling case for why democratizing drug discovery is critical, discussing how AI might help.

She shared a series of projects from her lab that demonstrate the capabilities and constraints of AI tools in this space. Among them was Paragraph, an antibody paratope prediction model that uses experimental and modeled structures to train an Equivariant Graph Neural Network. She also introduced AbLang, a language model for antibody sequences that revealed how skewed training data can bias model predictions toward germline residues, highlighting the importance of diverse and representative datasets.

For flexible protein modeling, Deane presented ITsFlexible, a method for predicting conformational flexibility in antibody CDR3 loops. This tool goes beyond static structure prediction, enabling more nuanced representations of protein dynamics.

Deane also examined AI tools for small molecule drug discovery, including structure-based docking and binding affinity prediction. She’s found that traditional docking methods often outperform newer AI-based models, especially when evaluated on physically realistic or out-of-distribution test sets. Benchmarks such as CASF-2016 were shown to have significant data overlap with training sets, raising questions about how well current AI models generalize. Her group proposed a zero-ligand bias benchmark as a more rigorous alternative.

Ultimately, Deane emphasized that while AI will undoubtedly continue to play a critical role in structure-based drug discovery, it is not a cure-all. Its success depends on thoughtful model design, realistic evaluation metrics, and—most importantly—robust, well-curated data.

 
 
 
 
 

Session Recaps
 
 

CSi

Wednesday, July 22: 

2025 marks the first year of Computational and Systems Immunology, CSI for short, COSI. We were excited to start the day with a packed room for Dr. Julio-Saez Rodriguez's Keynote talk on use of multiomic and spatial data alongside biological knowledge to model gene regulatory and signalling networks. This was followed by a talk from founder and former chair of the Computational Biology Department at Carnegie Mellon, Dr. Robert F. Murphy, on their vision transformer based cell annotation method for highly multiplexed tissue images. The morning session continued with talks on protein expression imputation from spatial transcriptomics data and spatial and single-cell study of the human thymus to understand how it becomes dysfunctional as we age.

The afternoon session started with a remote talk from Dr. Harinder Singh on deciphering transcriptional regulatory code of immune cells using an expanded DNA motif lexicon (such as composite TF motifs). Following talks by trainees looked at transcriptional regulators of Tfh differentiation, double negative B cells in cancer and prediction of pairing of heavy and light immunoglobulin chains. These were followed by talks on incorporation of phylogeny information for B cell epitope prediction, improvements for TCR-epitope binding prediction models and prediction of neoantigen immunogenicity in cancer. The last full talk of the afternoon introduced SHISMA, a novel tool to infer significant cell type-specific regulatory patterns with strong statistical guarantees. The day and the COSI session for this year was concluded with 5-minute short talks selected from submitted abstracts. See you next year in Washington DC to hear the latest and most exciting research directions in computational and systems immunology. Stay tuned for more CSI COSI activity including an online seminar series in the meantime!

 

Education

The Education COSI hosted a vibrant program covering a range of current issues in computational biology education. The session led off with an exciting keynote from Gabriela Rustici of AstraZeneca to share her experiences with bioinformatics training, emphasizing partnerships between industry and academia training and some of the challenges and opportunities that entails. This session also provides a bridge to the next day’s Workshop on Education in Bioinformatics (WEB), which will follow up on the theme of industry’s role in bioinformatics in education.

The program featured two proceedings talks by Alistair Kilpatrick and Pavlin G. Poličar and eight contributed abstract talks by Allissa Dillman, Ajay Mishra, Dusanka Nikolic, Aparna Nathan, Inimary Toby-Ogundeji, Amanda Saravia-Butler, Nadiia Kasianchuk, Nicola Mulder, and Alina Frolova. Collectively, these talks covered a wide range of ongoing and emerging topics in bioinformatics education, including computational and quantitative education of experimental biologists and public health workers, new pedagogy models and use cases, meeting challenges of content delivery in resource-limited environments, and the use of large language models (LLMs) to facilitate bioinformatics education.

Wrapping up the program was our closing keynote speaker, Patricia Carvajal-López of EMBL/EBI, sharing her work in developing and maintaining communities of practice in bioinformatics education and highlighting the recent success of the Bioinformatics Education Summit she co-organized in Mexico City this past May.

We thank the community for another great year in Bioinformatics Education and encourage participation in the WEB, visiting the Education posters, and contributing to community educational activities throughout the year.

 

HiTSeq

Today our amazing Hitseq track kicked off with cutting-edge advancements in algorithms for analyzing diverse types of high-throughput sequencing data.

We had two fantastic keynote speakers:

  • Valentina Boeva – “Learning variant effects on chromatin accessibility and 3D structure without matched Hi-C data”, and
  • Jonathan Goke – “Quantifying RNA Expression and Modifications using Long Read RNA-Seq”.

Our proceedings presentations featured:

  • A benchmark of deconvolution tools to estimate cell fractions in spatial transcriptomic data using ATAC- or RNA-based references, presented by Laura Mortens,
  • Oarfish – improving accuracy in long-read transcriptome quantification, presented by Zahra Zare,
  • Alevin-fry-atac – enabling frugal mapping of single-cell ATAC-seq data through virtual colors for precise genomic pseudoalignment, presented by Noor Pratap.

We also had exciting algorithm introductions including:

  • Icona – identifying structure-defining chromatic contacts, and
  • SplitSync – leveraging deep learning to improve splice site accuracy.

Finally, our sponsor PacBio presented “Bioinformatics analysis for long-read RNA sequencing: challenges and promises” by Elizabeth Tseng.

Follow us on our social media for more updates!

X @HiTSeq BlueSky @hitseq.bsky.social
 
 
 

NetBio

We wrapped up another engaging and well-attended half-day at the NetBio COSI, continuing the momentum from Day 1 with a packed room and lively discussion throughout the sessions. Today's program showcased a range of innovative work at the intersection of network
biology and translational medicine.

The morning opened with a proceedings talk by Dongmin Bang on
MixingDTA, a novel approach for improving drug-target affinity
prediction by extending Mixup with guilt-by-association principles. This
was followed by compelling presentations from Nicoleta Siminea,
Terence Egbelo, Johannes Kersting, and Samuele Firmani. These talks highlighted diverse computational strategies for supporting personalized therapies, addressing target-adverse event associations, building robust disease modules, and identifying antiviral
targets using message-passing neural networks.

After the lunch break, Piotr Sliwa presented a multilayer network
approach for identifying clinically relevant patient endotypes in
COVID-19 and sepsis, offering new insights into patient stratification.
We then turned our attention to visualization, with Chad Myers presenting the winners of the Cytoscape Visualization Competition - thank you to all who submitted their figures and voted! Congratulations to the winners Alejandra Paulina Perez Gonzalez, Michael Costanzo and Jong Chan Lim! Check out their submissions below!

The day concluded with a shared keynote talk with the Microbiome COSI by Fiona Brinkman, who walked us through the complexity of microbiome data and metadata management.

Many thanks to all NetBio speakers, poster presenters, moderators and organizers for their fantastic contributions!

If you haven't checked out the 45 NetBio posters yet, tomorrow's your last chance! Don't miss them!
 
 
 
 

SysMod

Wednesday, July 22

The 10th Annual SysMod Meeting was hoster by Matteo Barberis and Chiara Damiani, and it has been such a rich, dynamic, and truly international event. Whether through a talk, a poster presentation, joining the session, or simply listening and engaging in the discussions, the highly attendance of participants contributed to what felt like a very special milestone for the Systems Biology and Modelling community.

Three were the sessions covered in the 2025 edition of the meeting, with many highlights and takeways:

- Metabolic Modeling: New Variational Perspectives: The meeting opened with a session focused on metabolic modeling — a natural framework for bridging data-driven and mechanistic approaches. Ronan Fleming (University of Galway) shared his long-term vision to reformulate reaction kinetics using a variational framework, addressing a key limitation of genome-scale models: the lack of metabolite concentration dynamics. Talks by Lisa Corbeij, Garhima Arora, and Sultana Al Zubaidi explored new directions in dynamic and single-cell metabolic modeling.

- Multiscale Systems Biology and Control: Here we saw how computational models are being used to make sense of biological complexity across scales. Jasmin Fisher (UCL) delivered a powerful keynote on multiscale modeling in oncology, showing how intra- and inter-cellular signaling models can support data-driven and personalized medicine. Other talks featured reinforcement learning for optimizing JAK/STAT interventions (Nhung Duong), logic-based models for drug discovery (Bi-Rong Wang), and benchmarking of agent-based simulations (Arnau Montagud). The generative framework ARTEMIS (Sayali Alatkar) stood out for integrating autoencoders with Schrödinger Bridges to model gene expression dynamics from time-series single-cell data.

- Single-Cell Dynamics and Inference: The final session zoomed in on single-cell and spatial data. We saw high-resolution modeling of bacterial dynamics in microfluidic traps (Ati Ahmadi), metabolic atlas construction across organs (Erick Armingol), and deep learning approaches to spatiotemporal tissue dynamics (Koichiro Majima).

In addition to the three sessions, thre fast-paced talks added diversity and depth to the program: Jiachen Li (TFvelo: RNA velocity without splicing), Bingxin Lu (DNA repair and structural variation from cell-cycle modeling), and Marek Kimmel (Modeling telomere reconstruction).

Finally, so much energy and talent was spread in the poster session, with the poster prized being awarded to: Yanjun Liu (1st place), Rupinder Kaur (2nd place), and Sebastian Huß (3rd place). . Many Congratulations to the awardees! And many thanks to all the presenters for their valuable contributions.
 
 
 
 
 

🧭Career Compass Featured Job🧭
 
 

Postdoctoral Researcher (m/f/d) in Computational Oncology

University of Tübingen, Institute for Bioinformatics and Medical Informatics

Tübingen, Germany

KohlbacherLab.org

Job description:

The chair for Applied Bioinformatics at the University of Tübingen, Germany (Prof. Oliver Kohlbacher) is looking for a Postdoctoral Researcher position in Computational Oncology.
Position Details
Start Date: negotiable, available immediately
Duration: 3 years (initially)
Working Hours: Full-time (100%)
Salary: According to German salary scale TV-L E13

Description

The successful candidate will join an interdisciplinary team focused on applying bioinformatics to personalized oncology. The primary objective is to unravel molecular pathways of cancer by analyzing and integrating multi-omics and clinical data.

Key Responsibilities

  • Analyze and interpret multi-omics datasets, including genomic, transcriptomic, and clinical data from over 7,000 cases treated in molecular tumor boards.

  • Develop and implement bioinformatics pipelines to identify biomarkers and therapeutic targets.

  • Collaborate with clinicians and researchers to translate computational findings into clinical applications.

  • Publish research findings in peer-reviewed journals and present at international conferences.

  • Mentor junior researchers and contribute to the academic environment of the institute.

Opportunities

  • Access to a unique and comprehensive dataset combining genomic and clinical information from a large cohort of cancer patients.

  • Potential to establish an independent research group for candidates with a strong publication record and leadership potential.

  • Engagement with the University of Tübingen’s vibrant research community and participation in collaborative projects within the Clusters of Excellence ‘Machine Learning for Science’ and ‘Image-Guided and Functionally Instructed Tumor Therapies (iFIT)’.

Qualifications

Requirements

  • PhD in Bioinformatics, Computational Biology, or a related field.

  • Demonstrated expertise in analyzing high-throughput sequencing data and developing bioinformatics tools.

  • Proficiency in programming languages such as Python or R, and experience with relevant bioinformatics software.

  • Strong understanding of cancer biology and molecular oncology is advantageous.

  • Excellent communication skills and the ability to work collaboratively in a multidisciplinary team 

  • Previous experience in using machine learning methods will be beneficial
 
 
 
 

ISCB Values and EDI Initiatives
 
 

Have you seen the posters at the ISCB booth? We had two new ones created for the conference this year to share a reminder of the ISCB values, as well as the EDI Strategic Plan from 2025 to 2030. The EDI Strategic Plan highlights recent progress and outlines future goals to build a more inclusive, representative, and connected computational biology community. 

Check out a sneak peak of the posters below or visit the booth tomorrow!
 
 
 
 
 
 
 
 

❗Important Notes❗
 
 

POSTERS

All posters must be taken down by 4pm tomorrow, Thursday, July 24.

 

LUGGAGE AND COAT CHECK 

If you’re leaving right after the conference and your hotel is nearby, please ask the staff at the front desk of your hotel if they can hold your luggage for you. If your hotel isn’t nearby, you will be able to drop off your luggage at the cloakroom on the main floor of the ACC Galleria. 

NOTE: The cloakroom will close at 6:30pm on Thursday, July 24. If you do not collect your luggage by this time, anything left will be moved to the registration desk or given to lost and found.

**Neither ISCB event staff nor the ACC staff will be responsible for your luggage.

 

LOST ITEMS

If you’ve lost something while attending sessions during the conference, please visit the ISMB/ECCB registration desk as many lost items have been dropped off there. If your item isn't at the registration desk, please ask for any lost and found items at the ACC's Guest Relations desk.

 

TRANSPORTATION TO THE AIRPORT

Bus:If you're hoping to travel by bus, there are two main buses that service the Liverpool City Region:Arriva and Stagecoach. You can check the respective websites to find the correct bus line for you. Remember that the Liverpool ONE Bus Station is the closest service point for long distance coaches and the Liverpool ONE bus station is only a 5 minute walk from the ACC Liverpool.

Train: If you need to find a train to reach the city you're flying out from, Avanti West Coast has several cities they travel to or you can look at train options on Trainline.

Ride Share: If you want to travel by Uber to the airport, consider the ‘Share’ option to lower travel expenses, or connect with other conference attendees and split the fare for an Uber or taxi.

You can also try reviewing travel options on the Merseytravel website.
 
 
 
 
 

Coming Up Tomorrow, Thursday, July 24
 
 
 
 
 
 
 
 
 
 
 
 
 
 

ISCB On the Web

FlickrFacebookGoogleTwitterLinkedinYoutube