Bioinformatics Seminar

Title: Permutation Enhances the Rigor of Single-Cell Data Analysis

The online bioinformatics seminar is back this semester!

The first talk is on September 11, 2025, from 12:30-1:30 p.m. EDT. 

Zoom Link https://purdue-edu.zoom.us/j/97231496999

Speaker: Jingyi Jessica Li, Professor https://jsb-lab.org/people/jingyi-jessica-li/

Ensuring the reliability and accuracy of single-cell data analysis is critical, particularly in visualizing complex biological structures and addressing data sparsity. This talk introduces two novel statistical methods—scDEED and mcRigor—that leverage permutation-based techniques to enhance the rigor of these analyses. First, scDEED addresses the challenge of evaluating the reliability of two-dimensional (2D) embeddings produced by visualization methods like t-SNE and UMAP, which are commonly used to visualize cell clusters but can sometimes misrepresent data structure. scDEED calculates a reliability score for each cell embedding by comparing the consistency between a cell’s neighbors in the 2D embedding space and its pre-embedding neighbors, flagging dubious embeddings while confirming trustworthy ones. It also guides hyperparameter optimization for t-SNE and UMAP by minimizing dubious embeddings, thereby significantly improving visualization reliability across datasets. Second, mcRigor addresses the challenge of aggregating similar single cells into metacells, a common heuristic for sparse data that risks mixing dissimilar cells and introducing bias. mcRigor introduces a feature-correlation-based statistic to detect and filter heterogeneous metacells, improving metacell homogeneity, and provides a framework for optimizing partitioning algorithms by tuning hyperparameters. Additionally, mcRigor enables benchmarking and selecting the most suitable partitioning algorithm for a given dataset, ensuring more robust and reproducible insights in single-cell omics studies. Together, scDEED and mcRigor demonstrate the power of permutation-based approaches in refining single-cell data analysis, equipping researchers with tools for more accurate and trustworthy discoveries.

Bio: Jessica Li is Professor and Program Head of Biostatistics at the Fred Hutchinson Cancer Center, where she holds the Donald and Janet K. Guthrie Endowed Chair in Statistics, and an Affiliate Professor at the University of Washington. Her research develops reliable and interpretable statistical methods for analyzing complex biological data, with broad applications in gene regulation and biomedicine. Before joining Fred Hutch, she was a faculty member at UCLA from 2013 to 2025. Dr. Li’s contributions have been widely recognized, including the NSF CAREER Award, Sloan Fellowship, ISCB Overton Prize, COPSS Emerging Leaders Award, Guggenheim Fellowship, and Mortimer Spiegelman Award.

Host: Boran Gao (Stat/Bio)

Coordinator: Daisuke Kihara (Bio/CS), Majid Kazemian (Biochem/CS)

Past lectures are posted on the Youtube channel: https://www.youtube.com/channel/UCK4OqdS6xAxTCiVev843f6Q

(the channel includes only lectures approved by the speakers)