Multiomics analysis of cancer cell lines

cancer omics

DREAM challenge- Single cell signaling in breast cancer (genomics, transcriptomics, proteomics, phosphoproteomics) (link)

Biological motivation

This challenge involves analysis of the largest single cell signaling dataset (67 cancer cell lines). The motivation is to understand the heterogenous, time-dependent responses to multiple drug treatments. In understanding single-cell and cancer line responses to treatments, we aim to predict responses to drug candidates.

(link)

This analysis will be broken down into:

1. Exploratory data analysis: Examining time-dependent drug responses and identifying similarities in cell lines. (link)

2. Investigating feature importance in predicting phosphorylation changes in response to drug treatment using elastic net (ridge, LASSO) and tree-based models (random forests, xgboost) (link)

3. Prediction using neural networks (link)

4. Prediction using automl (tpot) with comparison to the above models.