List of participants and affiliations:
| Role | Name | Affiliation |
|---|---|---|
| Team Lead | Bingfang Ruth Xu | Frederick National Laboratory for Cancer Research (FNLCR), Leidos Biomedical Research, Inc., CMDL |
| Tech Lead | Daniel Sierra-Sosa | Assistant Professor, Department of Computer Science and IT, Hood College |
| Writer | Julie Bocetti | NICHD |
| Flex | Todd Young | Frederick National Laboratory for Cancer Research (FNLCR), Leidos Biomedical Research, Inc., CMDL |
| Flex | Brendan Reilly | Co-PI NSF 2221959, Co-PI NSF 2314315, Adjunct Lecturer Brooklyn College |
| Flex | Helga Saizonou | Tropical Infections Diseases Research Centre (TIDRC), Univeristy of Abomey-Calavi (UAC) |
Genomics of Drug Sensistivity in Cancer (GDSC) is a valuable resource for pharmagenenomic research. It has characterized 1000 human cancer cell lines and screened them with hundreds of drug compounds.
Currently, for a given drug, GDSC provides ANOVA test results for each of the 700 genomic features. We aim to identify a panel of features with a high predictive value for drug response of each individual drug using machine learning. We will first use over 700 genetic features as input and IC50 (Half-maximal inhibitory concentration) as the output in our model to predict drug response. We will test whether our panel of features includes the genomic features selected by the ANOVA test.
Data was generated from a 2016 paper: 'A Landscape of Pharmacogenomic Interactions in Cancer' (https://pubmed.ncbi.nlm.nih.gov/27397505/)
Data was accessed online from: https://www.cancerrxgene.org/downloads/drug_data (GDSC2 files)
We are using >700 genetic features that are binary, leading to a sparce distribution of information that is best modeled using decision trees. Our target variable (IC50) is continuous so we will be performing regression. To accomplish these 2 goals, we will use XGBoost.
This software was created as part of an NCBI codeathon, a hackathon-style event focused on rapid innovation. While we encourage you to explore and adapt this code, please be aware that NCBI does not provide ongoing support for it.
For general questions about NCBI software and tools, please visit: NCBI Contact Page