Projects

Home /
Projects

Bioinformatics Infrastructure

At Accent Therapeutics, I established a comprehensive bioinformatics infrastructure from the ground up to support all pre-clinical and clinical stage programs. This platform enables key activities such as target identification and evaluation, mechanism of action studies, disease biology research, translational medicine, and clinical trial design. Key features include:

Biomarker Discovery

One of my major responsibilities at Accent was to identify candidate predictive biomarkers for its oncology drugs. Since I joined Accent, I developed a biomarker discovery platform capable of handling various tasks to identify and explore potential biomarkers. The platform consists of an Accent proprietary database and a set of code that can efficiently generate biomarker reports with compound response data.

Comparative Genomics

Together with Kartik Shah from Love Lab at MIT, we assembled, annotated and compared three yeast strains. The study identifed major genomic rearrangements, a novel linear plasmid in K. phaffii, and 35 non-synonymous mutations in the industrial strain GS115. Deep RNA-Seq under diverse growth conditions revealed gene expression dynamics related to metabolism, secretion, and stress responses. Using co-expression analysis and functional enrichment, the study mapped regulatory networks and metabolic pathways. These comprehensive genomic and transcriptomic resources provide a solid foundation for improving strain engineering and optimizing recombinant protein production systems.

Copy Number Variants with Shallow Sequencing

While working at MIT, Kristin Knouse (Amon Lab) and I developed a workflow to reliably detect copy number variants (CNVs) from shallow whole-genome sequencing data. Our goal was to investigate previous reports of unexpectedly high levels of large-scale CNVs in somatic neural cells. Through simulations, we defined optimal parameters for CNV detection and found that, using these optimized parameters, the prevalence of large CNVs in somatic cells is lower than previously reported, which aligns better with biological expectations. This work established a standard for calling CNVs from shallow whole-genome sequencing data in single cells.

OLego, A Sensitive Splice Mapper

OLego is a program specifically designed for de novo spliced mapping of mRNA-seq reads. OLego adopts a multiple-seed-and-extend scheme and does not rely on a separate external mapper. It achieves high sensitivity of junction detection by using very small seeds (12–14 nt), efficiently mapped using the Burrows-Wheeler transform (BWT) and FM-index. This approach also makes it particularly sensitive for discovering small exons. OLego is implemented in C++ with full support for multithreading, enabling fast processing of large-scale data.

SpliceTrap, A Splicing Quantification Tool

SpliceTrap is a statistical tool for quantifying exon inclusion ratios in paired-end RNA-seq data, with broad applications in the study of alternative splicing. SpliceTrap estimates exon inclusion levels using a Bayesian inference approach. For each exon, it quantifies the extent to which it is included, skipped, or affected by size variations due to alternative 3’/5’ splice sites or intron retention. Additionally, SpliceTrap can quantify alternative splicing within a single cellular condition, without requiring a background set of reads.