I participated in the Open Problems - Single-Cell Perturbation competition hosted on Kaggle (https://www.kaggle.com/competitions/open-problems-single-cell-perturbations ), which aimed to predict gene expression responses to specific perturbations using single-cell data.

To address this challenge, I integrated multi-omic datasets, combining transcriptomic (scRNA-seq) and epigenomic (scATAC-seq) information. I employed Seurat for scRNA-seq analysis, Signac for scATAC-seq processing, and SCENIC+ to infer multi-omic gene regulatory networks (GRNs). Leveraging these layers of biological information, I developed multiple machine learning and deep learning models, incorporating GRNs as priors to enhance predictive performance.

This integrative approach enabled me to rank within the top 10% of over 1,000 participating teams.