Activity Number:
|
428
- Clustering and Dimension-Reduction Methods: From Omics to Single-Cell Data
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 10, 2022 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Statistics in Genomics and Genetics
|
Abstract #323230
|
|
Title:
|
Gene Embedding of Omic Data Using Gene Annotations
|
Author(s):
|
Zhexiao Lin* and Wei Sun
|
Companies:
|
University of Washington and Fred Hutchinson Cancer Research Center
|
Keywords:
|
dimension reduction;
graph neural networks;
single-cell omics
|
Abstract:
|
Dimension reduction of genes is an important intermediate step in analyzing single-cell omic data. Most existing methods, such Principal Component Analysis (PCA) or auto-encoder do not use gene annotation information. In this work, we proposed a new dimension reduction method by using protein-protein interaction from BioGRID and gene-gene similarity from gene ontology. The gene information was incorporated by using Graph Neural Networks for graph pooling. We applied the new gene embedding to unsupervised clustering and supervised cell type classification.
|
Authors who are presenting talks have a * after their name.