A Transformer for scATAC-scRNA Translation

scATAC-seq gives a comprehensive picture of the chromatin accessibility profile of a cell, covering not only protein-coding regions but also non-coding regulatory regions which are in theory missed by scRNA-seq. However, scATAC-seq data is highdimensional and noisy, aspects which when compounded with data scarcity present challenges for modeling on even seemingly-simple downstream tasks such as cell-type prediction. As such, researchers may benefit from access to a large library of models to evaluate. While we do not demonstrate state of the art results in any capacity, we provide an implementation of a simple representation of sparse tabular data that allows it to be inputted into the popular transformer family of architectures, and use this representation to train a transformer that predicts scRNA-seq given scATAC-seq. Our code is made available here https://github.com/rogershijin/GANOLI.

Name thesis:degree_name

Master

Department dc:contributor.department

Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2022

Author dc:creator

• Jin, Roger

Advisor dc:contributor.advisor

• Kellis, Manolis

Statement dc:rights

• In Copyright - Educational Use Permitted
• Copyright MIT

Licence dc:rights.uri

• http://rightsstatements.org/page/InC-EDU/1.0/

Handle dc:identifier.uri

https://hdl.handle.net/1721.1/147430

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/147430