RNA Therapeutics · Single-Cell RNA-Seq

Analysis of internal bulk and single-cell data to determine if compound was targeting a specific cell type with poorly defined marker genes, and investigate apparent differential signal in the two modalities.

Result: Identified a specific de novo subcluster of epithelial cells supporting compound activity in a particular region of the organ of interest; discoveries and slides presented to the board.

Microbiome Therapeutics · Metagenomics

Analyzed internal clinical and preclinical data in conjunction with exploratory meta-analysis of public metagenomics datasets to support publications and pre-IND reports. Partnered with internal team to roll out new data processing pipelines in the cloud.

Result: Abstracts and publications identified ICI response-associated taxa in multiple indications and compared these with robustly characterized product profile, to support further clinical development.

RNA Therapeutics · Oligo Target Prediction and Risk Assessment

Analyzed thousands of RNA-seq experiments to identify compounds with strong knockdown and low off-target risk. Developed an end-to-end platform for predicting oligo targets across species and a novel metric for assessing off-target risk, standardized across thousands of RNA-seq experiments.

Result: A fully integrated content management system for search and comparison of compounds and experiments across multiple programs and target areas.

RNA Therapeutics · Oligonucleotide Targets in Rare Disease

Analyzed hundreds of thousands of public RNA-seq experiments to identify unproductive RNA processing events (e.g., “poison” exons), which have the potential to be modulated by oligonucleotide compounds to upregulate gene targets in rare disease.

Result: A searchable database of RNA processing events identified in 425,000+ RNA-seq datasets suitable for target discovery.

Infrastructure · Data Wrangling

Migrated an ad-hoc, organic ecosystem of data and pipelines into one system. Ensured that pipelines were version-controlled, QCd, and reproducible, and also that any person can process, analyze, and store their data in a centralized repository.

Result: A fully custom platform for organizing existing and incoming data across the company. All pipelines were standardized, integrated, and accessible via the platform. Built a massively scalable backend in AWS to ensure no bottlenecks.