Application of ML models on genomic data

Abstract

To truly unlock the advantages of extensive datasets containing individual cell information, like the transcriptome and epigenome, it's crucial to incorporate them and look for ways to label each cell. However, it's a significant challenge to match one cell's epigenome profile with a vast collection of reference cells. That's where scEpiSearch comes in. It allows you to search, compare, and classify individual cells based on their open-chromatin profiles by comparing them to a large reference of both expression and open-chromatin datasets for single cells. ScEpiSearch aims to provide a platform to facilitate a primary step in single cell epigenome analysis. ScEpiSearch allows searching for matching cells using a single-cell epigenome profile. We can use our best efforts to optimize the ScEpiSearch so it can be run on a normal laptop. The goal is to brainstorm different techniques of Machine Learning to possibly create a lite version of ScEpiSearch which can run on any OS without any glitch , but to do so we have to understand the basic concepts of the software and the way it works technically using hashing techniques , knowledge of the techstack of the software involved and etc.