Cole Deisseroth
Medical Scientist Training Program (MD/PhD)
Positions
- Medical Scientist Training Program (MD/PhD)
-
ÌÇÐÄÊÓƵ of Medicine
Houston, TX, US
- GS3
-
Genetics & Genomics
ÌÇÐÄÊÓƵ of Medicine
Houston, Texas, United States
Mentor: Huda Zoghbi, MD Mentor: Zhandong Liu, PhD
Education
- BS from Stanford University
- 06/2019 - Stanford, California, United States
- Computer Science
Professional Interests
Professional Statement
Most of my research experience is in automatically extracting discrete, machine-readable information from unstructured text, and using that information to fuel accurate downstream analyses. In my undergraduate training, I worked in the lab of Dr. Gill Bejerano, and led the development of ClinPhen, a tool that automatically identifies and encodes phenotypic information in clinical notes. The extracted phenotypic data can be fed directly to phenotype-matching tools designed to aid the diagnosis of Mendelian disorders (PMID 30514889).
In my post-baccalaureate work (and continuing into my PhD training), I developed another tool designed to identify known pharmacologic gene regulators and predict new ones. This tool is called PARsing ModifiErS via Article aNnotations (PARMESAN, PMID 37741276), and it automatically extracts discrete claims of gene-gene and drug-gene regulatory relationships. It then establishes an overall consensus for each relationship, and prioritizes the relations based on the amount of supporting and opposing evidence. Lastly, it feeds the relations into two-step pathway analyses to identify potential indirect effects of one molecular entity on another. Higher-scoring predictions are more likely to show the correct directionality (up- versus down-regulation) of known regulatory relationships. However, one of my current goals is to evaluate its effectiveness in aiding discovery efforts, by using it to guide in vitro regulator screens.
In my post-baccalaureate work (and continuing into my PhD training), I developed another tool designed to identify known pharmacologic gene regulators and predict new ones. This tool is called PARsing ModifiErS via Article aNnotations (PARMESAN, PMID 37741276), and it automatically extracts discrete claims of gene-gene and drug-gene regulatory relationships. It then establishes an overall consensus for each relationship, and prioritizes the relations based on the amount of supporting and opposing evidence. Lastly, it feeds the relations into two-step pathway analyses to identify potential indirect effects of one molecular entity on another. Higher-scoring predictions are more likely to show the correct directionality (up- versus down-regulation) of known regulatory relationships. However, one of my current goals is to evaluate its effectiveness in aiding discovery efforts, by using it to guide in vitro regulator screens.
Selected Publications
-
" ;
Pubmed PMID: . -
Cole A Deisseroth, Won-Seok Lee, Jiyoen Kim, Hyun-Hwan Jeong, Ryan S Dhindsa, Julia Wang, Huda Y Zoghbi, Zhandong Liu. " " 2023 Oct 5;
Pubmed PMID: .
Skills
- Coding in Bash and Python; applying artificial intelligence methods for natural language processing
to edit your profile