Skip to content

Oct - Khanh Dao Duc

Event Details

Date/Time:

Thursday, October 19th, 2023 11:00am - 1:30pm PT

Location:

UBC, Michael Smith Laboratories, MSL 102

SFU Big Data Hub, room ASB 10900 (Live Stream Location)

RSVP:

If you are interested in attending this seminar in person, please fill out the RSVP form.

Featured Speaker: Dr. Khanh Dao Duc

Affiliations:

  • Assistant Professor, Mathematics, University of British Columbia
  • Associate Member, Computer Science, University of British Columbia
  • Associate Member, Zoology, University of British Columbia

Talk Title: Bioinformatics pipelines, algorithms and non-linear metrics for quantifying heterogeneity in large biological datasets

Abstract:

Recent advances in sequencing and imaging technologies have brought our understanding of biological processes and structures at unprecedented details and scales, but also require new tools to process and interpret large and complex datasets. In this context, I will present our effort to analyze produce various end-to-end pipelines, databases and softwares that combine bioinformatics with mathematical and machine learning methods. One prime application will be to investigate the structural and functional heterogeneity of the ribosome from structural data. For that purpose, my group recently developed a database that provides a standardized nomenclature for comparing ribosomal components (proteins, RNA, ligands) across all the available structures from the PDB, with several specialized visualization tools. I will present our recent effort to further develop this tool and discuss applications in the context of the ribosome exit tunnel, a key compartment contains the nascent polypeptide chain. In a second part, I will describe how this work inspired us to develop more general methods, that use recent advances in computational geometry and machine learning for cryo-EM and biological shape analysis, from molecules to cells. These methods will include various applications of optimal transport-based distances for comparing 3D voxelized maps that outperform standard linear methods in Cryo-EM, and Riemannian elastic metrics that also yield better clustering and separation in comparison with the linear metric on various datasets of 2D cancer cell images.

Bio:

Dr. Dao Duc got his PhD in applied mathematics in 2013 from the Ecole Normale Superieure under the supervision of David Holcman (Paris, France), where he studied stochastic models in biology. From 2014 to 2019, he was a Postdoctoral fellow in Professor Yun Song's lab at UC Berkeley (2014-2015 and 2018-19) and UPenn (Simons Math+X postdoctoral fellow), where he mainly investigated the determinants of translation dynamics by combining new mathematical and computational approaches with Ribo-Seq and Cryo-EM data. Since 2019, he has joined UBC as an Assistant Professor in the Mathematics Department and associate member of the Computer Science Department. His lab focuses on using mathematical and computational methods to model biological processes and analyze complex biological datasets, with multiple collaborations across different departments at UBC.

Trainee Speaker: Courtney Hoskinson

Affiliation: Turvey Lab

Talk Title: Omics in the early-life gut: delayed gut microbiota maturation as a hallmark of pediatric allergic disease