Welcome to PRESGENE

Essential gene prediction helps to find minimal genes indispensable for appropriate cellular function and survival of any organism. Machine learning (ML) algorithms have been useful for prediction of gene essentiality and annotation. However, currently available ML pipelines perform poorly for organisms with limited experimental annotation of genes. The major objective of this work is the development of a new machine learning pipeline to help in annotation of essential genes of less explored disease-causing organisms for which very limited experimental data is available.

The proposed strategy combines unsupervised feature selection technique, followed by dimension reduction using Kamada-Kawai algorithm and semi-supervised machine learning algorithm employing Laplacian Support Vector Machine (LapSVM), for observing a distinct pattern in clustering, classification, and prediction of essential and non-essential genes from genome-scale metabolic networks using very limited labelled dataset. A novel scoring technique SSMSS, equivalent to MCC and auROC, has been proposed for selection of best model when supervised performance metrics calculation is difficult due to lack of data.

The unsupervised feature selection technique select relevant features and dimension reduction step helped to project the high dimensional data on a 2-D plane for observing a distinct pattern in clustering of essential and non-essential genes as an arc of a circular layout. LapSVM then created a curve that dissected this circle for the classification and prediction of essential genes with high accuracy (auROC > 0.85) even with 1% labelled data for model training.

Using graph-based semi-supervised machine learning scheme, a novel integrative approach has been proposed for essential gene prediction that shows universality in application to both Prokaryotes and Eukaryotes with limited labeled data. The essential genes predicted using the pipeline provide an important lead for the prediction of gene essentiality and identification of novel therapeutic targets for antibiotic and vaccine development against disease-causing parasites.

For Details See our Publication:
Sutanu Nandi, Piyali Ganguli And Ram Rup Sarkar*, Essential gene prediction using limited gene essentiality information–An integrative semisupervised machine learning strategy, PLoS ONE., 15(11), e0242943 (2020), DOI:10.1371/journal.pone.0242943.