Nag, S. (BMEB) – Personalized Diploid Genome Graphs for Accurate Somatic Variant Discovery

Many somatic variant-calling pipelines begin by aligning tumor and matched-normal sequencing reads to a single linear reference genome, such as GRCh38. Because every individual differs substantially from this reference, this approach can introduce reference bias, causing reads to map incorrectly or not at all and potentially leading to missed somatic variants or germline variants being misclassified as somatic. I propose replacing the generic reference with a personalized diploid genome graph constructed from a donor-specific assembly (DSA), which represents both inherited haplotypes of the individual. I will develop this framework by (1) generating haplotype-resolved, telomere-to-telomere DSAs for cancer reference cell lines, (2) developing haplotype-aware graph alignment and adapting DeepSomatic to call variants against personalized diploid genomes, and (3) applying this approach across tissues from SMaHT donors to improve somatic mosaicism detection and characterize shared and tissue-specific mutations.

Event Host: Sagorika Nag, Ph.D. Student, Biomolecular Engineering & Bioinformatics 

Advisor: Benedict Paten

Zoom: https://ucsc.zoom.us/j/99844148597?pwd=amp5Nhmj2UeodTADUdaJwZsKtscRMG.1

Passcode: 685655