• Daniel E. Platt, IBM Research, NY, United States
• Aritra Bose, IBM Research, NY, United States
• Kahn Rhrissorrakrai, IBM Research, NY, United States
• Chaya Levovitz, IBM Research, NY, United States
• Laxmi Parida, IBM Research, NY, United States

The emergence of COVID-19 created incredible worldwide challenges but offers unique opportunities to understand the physiology of its risk factors and their interactions with complex disease conditions, such as metabolic syndrome. To address the challenges of discovering clinically relevant interactions, we employed a unique approach for epidemiological analysis powered by Redescription-based TDA (RTDA). Here RTDA was applied to Explorys data to discover associations among severe COVID19 and metabolic syndrome. This approach was able to further explore the probative value of drug prescriptions to capture the involvement of RAAS and hypertension with COVID-19, as well as modification of risk factor impact by hyperlipidemia on severe COVID-19. RTDA found higher-order relationships between RAAS pathway and severe COVID-19 along with demographic variables of age, gender, and comorbidities such as obesity, statin prescriptions, hyperlipidemia, chronic kidney failure and disproportionately affecting African Americans. RTDA combined with CuNA (Cumulant-based Network Analysis) yielded an higher-order interaction network derived from cumulants that furthered supported the central role that RAAS plays. TDA techniques can provide a novel outlook beyond typical logistic regressions in epidemiology. From an observational cohort of electronic medical records, it can find out how RAAS drugs interact with comorbidities, such as hypertension and hyperlipidemia, of patients with severe bouts of COVID-19. Where single variable association tests with outcome can struggle, TDA's higher-order interaction network between different variables enables the discovery of the comorbidities of a disease such as COVID-19 work in concert.

• George Acquaah-Mensah, Massachusetts College of Pharmacy & Health Sciences, United States
• Boris Aguilar, Institute for Systems Biology, United States
• Kawther Abdilleh, Pancreatic Cancer Action Network, United States
• Ronald Taylor, National Cancer Institute, United States

Breast cancer is among the deadliest cancers for women in the world. Breast cancer has four distinct molecular subtypes which are determined by gene expression profiling. There are observed differences between the races at the molecular subtype level. Moreover, racial disparity exists between younger White and Black breast cancer patients as it relates to survival and recurrence. Differences have been characterized using molecular data. Imaging data are also valuable in breast cancer diagnostics and can be used to identify these differences as well. In this study, we applied ML techniques to identify molecular subtypes from MRI data deriving from a single-institutional, retrospective collection of 922 biopsy-confirmed invasive breast cancer patients that were collected over a decade. We limited our analyses to Black or White patients who were 50 years or younger at diagnosis (n=346).
RandomForest and AdaBoostM1 in WeKa were applied to over 500 MRI features. We found that imaging features alone or in combination with TCGA gene expression data were predictive of molecular subtype and recurrence events for both racial groups. As an example, the most predictive imaging features in the Breast Fibroglandular Tissue Volume imaging category for Black patients was BreastVol and for White patients was TissueVol_PostCon.Notably, these two imaging features were significantly associated with one race compared to the other. These results suggest that radiomic imaging data can be used to predict breast cancer recurrence and molecular subtype and can have an impact on clinical outcomes.

• Eric Glynn, Princeton University, United States
• Dario Ghersi, University of Omaha, United States
• Mona Singh, Princeton University, United States

Deep learning tools that predict peptide binding by major histocompatibility complex (MHC) proteins play an essential role in developing personalized cancer immunotherapies and vaccines. In order to ensure equitable health outcomes from their application, MHC binding prediction methods must work well across the vast landscape of MHC alleles. Here we show that there are alarming differences across individuals in different racial and ethnic groups in how much binding data are associated with their MHC alleles. We introduce a machine learning framework to assess the impact of this data disparity for predicting binding for any given MHC allele, and apply it to develop a state-of-the-art MHC binding prediction model that additionally provides per-allele performance estimates. We demonstrate that our MHC binding model successfully mitigates much of the data disparities observed across racial groups. To address remaining inequities, we devise an algorithmic strategy for targeted data collection. Our work lays the foundation for further development of equitable MHC binding models for use in personalized immunotherapies.

• Selene L Fernandez-Valverde, The University of New South Wales, Australia
• Ian Willis, EMBL-EBI, United Kingdom
• Piraveen Gopalasingam, EMBL-EBI, United Kingdom
• Marco Cristancho, Consorcio Universidades de Manizales, Colombia
• Cath Brooksbank, EMBL-EBI, United Kingdom
• Rebeca Campos-Sánchez, Universidad de Costa Rica, Costa Rica

The CABANA project (Capacity Building for Bioinformatics in Latin America), funded by the UK's Global Challenges Research Fund from 2017 to 2022, aimed to strengthen bioinformatics capacity and broaden applications across Latin America. Focusing on three change areas: communicable diseases, sustainable food production, and protection of biodiversity, the five-year initiative executed a comprehensive array of activities spanning ten countries (9 in Latin America in addition to the UK). These included data analysis workshops, train-the-trainer programs, secondments, eLearning development, knowledge exchange forums, and collaborative research endeavours. CABANA proved highly successful, accomplishing all objectives while generating a substantial regional impact. The project pioneered a novel approach where research needs dictated the training provided. Its enduring legacy encompasses over 800 trainees and multiple scholarly publications.

Its continuation, CABANAnet (funded by the Chan Zuckerberg Initiative), is building upon this foundation. Premised on computational biology, CABANAnet's capacity-building initiatives will feature research collaborations, workshops, a Train-the-Trainer curriculum, internships, eLearning, knowledge-sharing events, website development, and publications. Synergistic collaborations with complementary bioinformatics efforts such as SoiBio, ISCB, Global Data Alliance, and others will be encouraged. Through this multifaceted program, we aspire to foster positive change across local policies, economies, and scientific advancements throughout the region.