GitHub - xmkrohannon/Machine-Learning: Accurate prediction of the postmortem interval (PMI) is one of the most important and complex tasks performed by forensic coroners. Transcriptomics offers a better alternative for PMI estimation as differential transcript expression within various tissues can provide precise indicators and ultimately make better predictions, accounting for the tissue types, forensic contexts, and other confounding factors. The Genotype-Tissue Expression (GTEx) Project provides RNA sequencing data. Here in this study, we have utilized a dataset from the GTEx project (dbGaP accession number: phs000424.v7.p2) which was generated using 9,777 tissue samples from 550 subjects for developing PMI prediction models. We are mining this data to further enhance key predictors associated with PMI using blood and salivary gland tissue types. Several machine-learning techniques have been trained and tested using the dataset to find the most useful and practical model to obtain the optimum prediction accuracy. Although the precision of data is slightly compromised, a categorical approach offers flexibility in prediction by widening the PMI period. This novel study will immensely aid forensic investigators in PMI prediction and can contribute to a broader understanding of postmortem biological processes. Furthermore, validation of the model with biological assays has been performed on other cohorts to evaluate prediction performance.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
Blood_SignifGeneTPMs_Covariates.txt		Blood_SignifGeneTPMs_Covariates.txt
GroupB_MethodologyReport.docx		GroupB_MethodologyReport.docx
Machine Learning Models Full File.py		Machine Learning Models Full File.py
Machine Learning Models Top Genes.py		Machine Learning Models Top Genes.py
topgenes.csv		topgenes.csv

About

Accurate prediction of the postmortem interval (PMI) is one of the most important and complex tasks performed by forensic coroners. Transcriptomics offers a better alternative for PMI estimation as differential transcript expression within various tissues can provide precise indicators and ultimately make better predictions, accounting for the t…