.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.png)
Minisymposium
MS4H - Synergizing AI and HPC for Pandemic Preparedness with Genomics and Clinical Risk Assessment
Replay
Session Chair
Description
To address emerging virus variants, our strategy integrates next-gen vaccines and personalized disease treatments. In Bioinformatics Sequencing, HPC and AI power vaccine development and infection control. Simultaneously, AI-Driven Clinical Risk Assessment aids healthcare in pandemics. Personalized disease stratification involves AI models for risk assessment and interpretability-guided deep learning in medical applications. Standardizing EHR and Federated Learning ensures data integrity and privacy. In Bioinformatics Sequencing, we tackle challenges through: Drug Discovery for Next-Gen Vaccines: Applying bioinformatics to identify therapeutic candidates from genomic data for infectious diseases. Evolutionary Analysis for Infection Spread: Analyzing viral sequence data to identify important genes, functions, and evolution for minimizing and tracking infection spread. Accelerating Genotype-Phenotype Workflow: Correlating genotype to phenotype for efficient drug discovery in functional genomics. For AI-Driven Clinical Risk Assessment, methods include: AI Models for Disease Progression: Using advanced deep learning to characterize disease subtypes based on unsupervised and supervised learning. Interpretability-Guided Deep Learning: Enhancing comprehension in medical AI by addressing bias, shortcut learning, and susceptibility to attacks. Standardizing EHR and Federated Learning: Ensuring uniform Electronic Health Records (EHRs) usage, standardizing data formats, and addressing privacy concerns through federated learning. This minisymposium brings together experts to accelerate pandemic preparedness with clinico-genomic-data to improve diagnosis.
Presentations
This talk outlines the development of AI models for Risk Assessment and Disease Progression in Pandemic Preparedness. Leveraging advanced deep learning techniques, these models integrate diverse data sources like images, clinical features, and laboratory data to characterize disease subtypes and personalize risk assessment. Such AI-driven Clinical Risk Assessment aids healthcare professionals in resource allocation, especially in high-demand scenarios such as the COVID-19 pandemic.The talk will further demonstrate a specific AI model from an SNSF-funded research project focused on the development of an AI-multiomics-based prognostic stratification of acute and chronic COVID-19 patients. Using chest CT scans and clinical data, two models, AssessNet-19 and ChronRisk-19, significantly enhance severity assessment and long-COVID prediction. Evaluated on diverse patient cohorts, the models outperform radiologists and single-class lesion models. The research underscores the potential of AI in improving COVID-19 patient care, offering insights for future clinical workflows.
SARS-CoV-2 Omicron variants BA.2.86 and JN.1 have mutations that have raised concerns over their health impact. Genomic surveillance of JN.1 has shown it to be the dominant variant circulating in the USA.
Empirical studies on immune evasion and transmissibility on BA.2.86 and JN.1 are contradictory. To assess immune evasion of BA.2.86 and JN.1, we performed in silico analyses of the Receptor Binding Domain (RBD) of SARS-CoV-2 variants. We calculated the relative binding affinity of neutralizing antibodies derived from vaccinated patients, infected patients, and therapy to the RBDs. To assess transmissibility, we calculated the relative binding affinity of the RBDs to the Angiotensin Converting Enzyme-2 (ACE2) host cell receptor.
We found minor changes in some binding affinity metrics for neutralizing antibodies and ACE2 to RBDs of BA.2.86 and JN.1. However, most changes are not statistically significant. We conclude that BA.2.86 and JN.1 have inconsequential changes in immune evasion or transmissibility. We caution that genomic surveillance that counts mutations and the prevalence of a variant does not reveal the functional and health impacts of the variant. In concordance with our results, there has not been a surge on hospitals in the USA with the rise in the prevalence of BA.2.86 and JN.1.
Efficient analysis of Big Data presents unique challenges across molecular biology, genetics, biomedical sciences, and healthcare, particularly in advancing personalized diagnostics and therapeutics. This necessitates innovative strategies for effective information management. Here, we explore the evolving landscape of Bioinformatics and Computational Biology, reshaping data storage, management, and access through advancements in High-Performance Computing (HPC) and Big Data. Our project aims to develop sustainable scientific gateways for bioinformatics by integrating HPC environments, scientific workflows, and artificial intelligence. The BioinfoPortal (https://bioinfo.lncc.br/) serves as a multiuser infrastructure for bioinformatics applications, seamlessly connected to the HPC infrastructure of Santos Dumont (SDumont, https://sdumont.lncc.br/), the largest supercomputer in Latin America, boasting 5.1 Petaflops and 36,472 computational cores across 1,134 computational nodes. We discuss challenges in executing applications efficiently and provide insights to enhance computational resource utilization, exploring large-scale bioinformatics experiments for potential improvements. Currently, we are integrating artificial intelligence techniques to predict the optimal use of SDumont computational resources.
The German Human Genome-Phenome Archive (GHGA) is dedicated to addressing the challenges of managing human genomics and phenotypic data. As part of the German National Research Data Infrastructure (NFDI), it connects German researchers to the global genome research landscape, in collaboration with initiatives like the European Genome-Phenome Archive (EGA) and the European Genomic Data Infrastructure (GDI). It will enable the integration of data into and from national and international studies while safeguarding sensitive patient information. GHGA follows an approach that involves establishing secure IT infrastructure for omics data management, implementing an ethico-legal framework for data protection, harmonizing metadata schemas, and standardizing omics data processing workflows, while embracing Global Alliance for Genomics and Health (GA4GH) standards. Beyond establishing a data repository, GHGA actively engages in the development, standardization, and benchmarking of bioinformatics workflows, collaborating with the nf-core and NCbench communities. The recently launched GHGA Metadata Catalog facilitates the discovery of non-personal metadata, laying the foundation for the upcoming GHGA Archive. GHGA will enable cross-project analysis and promote new collaborations and research projects in the international context of genome research. This will also be of high importance for new national projects such as the upcoming genomDE project within Germany.