Development and improvement of research software for DNA methylation and RNA sequencing analysis

The National Institutes of Health Office of the Director awarded the Van Andel Research Institute $190,764 on September 1, 2026, under the Trans-NIH Research Support program (CFDA 93.310) to develop and improve research software for DNA methylation and RNA sequencing analysis.

The award funds a research software engineer in Dr. Hui Shen's laboratory to enhance existing bioinformatics tools. Specific deliverables include removing unused code from BISCUIT (a whole-genome bisulfite sequencing analysis tool), improving the speed of DNA methylation extraction from aligned BAM files, and creating a testing framework. The engineer will simplify DUPSIFTER by recoding how bisulfite strands are delineated and add functions for differentiating unique molecular indexes in duplicate read signatures. Additional work will improve RNA-seq analysis tools. The existing software suite—including BISCUIT, BISCUITEER, DUPSIFTER, and BISPLOTTI—has achieved significant adoption, with BISCUIT and DUPSIFTER combined downloaded more than 23,700 times via Bioconda and BISCUITEER downloaded more than 9,000 times via Bioconductor from more than 2,700 unique IP addresses.

Performance occurs at the Van Andel Research Institute in Michigan from September 1, 2026, through August 31, 2029. The assistance type is a Project Grant.

Abstract

I am a research software engineer in DR. HUI shen's laboratory at the van andel institute (VAI), where we focus on the origin of cancer through DNA methylation and single-cell technologies. I am currently supported by DR. Shen's ESI merit award (R37CA230748) and a cooperative agreement for the genomic data analysis center's integrative cancer epigenomic data analysis center (U24CA264023) from the NCI, as well as a commons fund award (UM1DA058219). As part of these efforts, i develop standards-compliant tools that work across a wide spectrum of inputs. For example, i developed or CO-developed biscuit, biscuiteer, dupsifter, and bisplotti - tools for analyzing whole genome bisulfite sequencing (wgbs) and wgbs-like data. Biscuit and dupsifter are available on bioconda and have been downloaded a combined >23,700 times, while biscuiteer is available on bioconductor and has been downloaded >9,000 times by >2,700 unique IP addresses. I worked with the original storm-SEQ technology developers (at VAI and takara BIO) and developed a quality control (QC) pipeline for detecting and mitigating contamination from adapter oligonucleotides, missed plate wells during cell sorting, and failed ribo-depletion. In addition, i developed a pipeline for adding an in silico barcode to each read and aligning storm-SEQ data to a reference genome. All my tools are well-documented in detailed github repository readme files, r markdown vignettes, or static websites. In this R50 project, i will support DR. Shen's projects by removing unused code from biscuit, improving the speed which DNA methylation information is extracted from aligned bams, and creating a testing framework. I will simplify dupsifter by RE-coding how the two bisulfite strands are delineated and add functions for differentiating unique molecular indexes (umis) in the duplicate read signature. For the RNA-SEQ tools, i will add new storm-SEQ QC metrics and a function for including cell barcodes and umis simultaneously. For my independent R50 research projects, i will develop new tools for analyzing thousands of wgbs samples across multiple tissues from the sequencing read archive (SRA) and identifying conserved hyper- and hypo-methylated regions. I will also develop a tool that reads raw fastq files and filters out potential strand invasion events that occur when using template switching oligonucleotides during reverse transcription or RNA-RNA interaction events. These chimeric reads can affect downstream analysis through improper collapsing of umis or misidentifying RNA-RNA interactions as structural variants. Through this R50 award, i will complete my MS in software engineering; grow my understanding of data compression techniques; apply that knowledge to store and manipulate large biomedical datasets; and apply formal software engineering principles to the tools that i develop, thus establishing myself as a high-quality research software engineer at VAI, and for the larger research software field.