Keynote Speakers:
Keynote Speaker (confirmed):
Kishwar Shafin, Ph.D.
Creating the Next Generation of Genome Analysis Tools with Deep Learning
9:00AM Monday 23 March 2026.
Abstract: Accurate genome analysis is fundamental to modern genomics, yet it continues to face significant obstacles. In germline variant calling, reference bias and sequencing errors remain persistent challenges. This talk presents a suite of deep learning-based tools designed to address these limitations. We first discuss DeepVariant, a convolutional neural network framework that has set new standards for germline variant calling accuracy. Building on this, we introduce Pangenome-aware DeepVariant, which leverages the Human Pangenome Reference Consortium (HPRC) draft to mitigate reference bias inherent in linear reference genomes, thereby improving mapping and variant calling performance.
To support the construction of high-quality pangenomes, we present DeepPolisher, a tool developed to correct base-level errors in genome assemblies. Finally, we detail DeepSomatic, a somatic variant caller trained on extensive datasets to identify tumor-specific variants with high precision. Together, these tools represent a comprehensive deep learning ecosystem for advancing germline, assembly, and cancer genomics.
Bio: Dr. Kishwar Shafin is a Senior Research Scientist at Google and the Technical Lead for DeepVariant. His research sits at the intersection of artificial intelligence and genomics, focusing on developing deep learning methods to improve the accuracy of variant calling and genome assembly. He has led the development of several state-of-the-art tools, including DeepVariant, DeepSomatic, and DeepPolisher. A key contributor to the Telomere-to-Telomere (T2T) project and the Human Pangenome Reference Consortium, his work focuses on making genomics more comprehensive and equitable. Before joining Google, he completed his PhD at the UC Santa Cruz Genomics Institute, where he was part of the team that set a Guinness World Record for the fastest DNA sequencing technique...
Linkedin: https://www.linkedin.com/in/kishwarshafin/
https://scholar.google.com/citations?user=QtsS4x8AAAAJ&hl=en