Abstract
Current single-cell spatial transcriptomics (sc-ST) technologies are limited by the number of genes they can simultaneously profile. Being based on fluorescence in situ hybridization, they are typically limited to panels of about a thousand genes. This limitation prevents researchers from exploring the full complexity of biological systems, such as cell-cell communication mediated by ligand-receptor interactions or the activation of specific signaling pathways. We propose scGIST, a constrained feature selection tool that designs sc-ST panels prioritizing user-specified genes without compromising cell type detection accuracy. We demonstrate scGIST’s efficacy in diverse use cases, highlighting it as a valuable addition to sc-ST’s algorithmic toolbox. scGIST enables researchers to move beyond identifying cell types and uncover novel biological insights by including genes that are often overlooked in traditional panel designs.
Abul Hassan Samee, Ph.D.
Assistant Professor of Integrative Physiology department
Dr. Samee is an Assistant Professor in the Integrative Physiology department of Baylor College of Medicine. His group is interested in machine learning (ML), single-cell, and spatial omics. In particular, they leverage machine learning, algorithms, and biophysics to develop ML models for massive and complex biological datasets. Current projects include single-cell and spatial omics data modeling for cardiac regeneration, Alzheimer’s, and cancer.
