ptellier / covid-smoker-ml

Prediction of smoker status of individual cells from single-cell RNA sequencing data. Data visualizations of important features and various model performance metrics

Gene expression levels from scRNA-Seq between smokers and non-smokers in COVID-19 Patients

Abstract

Muus et al. (2020) investigated the correlation between ACE2, TMPRSS2, and CTSL expression and smoking in COVID-19 patients. As an extension to this study, this paper investigated the predictability of smoking habits using scRNA-Seq gene expressions of COVID-19 patients. Topic modelling, logistic regression, and random forests revealed that CTSB, CTSC, CTSL may be good predictors of smoking behaviours. It is recommended that the authors of the original study investigate other potential genes’ expressions with smoking in COVID-19 patients and not limit their analysis to ACE2, TMPRSS2, and CTSL.

Keywords

Machine Learning, Statistics, Topic Modelling, Random Forest, Logistic Regression, Single-cell RNA, Sequencing, Non-negative Matrix Factorization, COVID-19, Smoking

Full Report

Read full PDF