Venky Soundararajan

This paper develops a deep learning model to predict neutralizing activity of COVID-19 therapeutics and vaccines against SARS-CoV-2 variants, finding reduced activity against newer variants like EG.5 and XBB.1.16.

The paper identifies SARS-CoV-2 peptides similar to human cardiac proteins, suggesting potential molecular mimicry that may contribute to cardiac inflammation in COVID-19.

The paper analyzes mutations across the SARS-CoV-2 proteome in the Delta variant from 2 million genomes, finding high diversity and specific mutations in non-Spike proteins compared to other variants.

This study uses a literature-derived knowledge graph to improve scRNA-seq annotation by leveraging gene-cell type associations from biomedical texts.

The paper uses a data science platform to integrate biomedical text and single-cell RNA-seq data to identify new SARS-CoV-2 receptor expression patterns and potential infection targets.

The paper identifies a SARS-CoV-2 peptide that mimics a human ENaC cleavage site, exploiting host proteolytic machinery and linking it to COVID-19 pathophysiology.

The paper introduces a platform for real-time synthesis of biomedical knowledge using unsupervised neural networks, integrating unstructured literature with structured databases.