Identifying essential genes across eukaryotes by machine learning

Abstract

Identifying essential genes on a genome scale is resource intensive and has been performed for only a few eukaryotes. For less studied organisms essentiality might be predicted by gene homology. However, this approach cannot be applied to non-conserved genes. Additionally, divergent essentiality information is obtained from studying single cells or whole, multi-cellular organisms, and particularly when derived from human cell line screens and human population studies. We employed machine learning across six model eukaryotes and 60,381 genes, using 41,635 features derived from sequence, gene functions and network topology. Within a leave-one-organism-out cross-validation, the classifiers showed a high generalizability with an average accuracy close to 80% in the left-out species. As a case study, we applied the method to Tribolium castaneum and validated predictions experimentally yielding similar performance. Finally, using the classifier based on the studied model organisms enabled linking the essentiality information of human cell line screens and population studies.

Genome-wide functional screens enable the prediction of high activity CRISPR-Cas9 and -Cas12a guides in Yarrowia lipolytica

“…While this work focuses on the development of DeepGuide for its specific use in Y. lipolytica, the same experimentalcomputational workflow that involves (i) library design, (ii) generating genome-wide guide activity profiles, (iii) predictor design (learning and optimization), and (iv) external validation, can be readily applied to other fungal species, broadly to prokaryotes, and any other organisms in which genome-wide functional screens can be used to estimate sgRNA activities. Moreover, DeepGuide adds to the growing number of examples in which deep learning is being used to solve complex problems in molecular biology, e.g., the prediction of essential genes 40,41 . BatchNormalization layers (see Supplementary The encoder in the second network has the same structure as the encoder in the CAE (see Supplementary Table 3).…”

Machine learning on large scale perturbation screens for SARS-CoV-2 host factors identifies β-catenin/CBP inhibitor PRI-724 as a potent antiviral

“…The procedure for feature generation was similar as published earlier for essential gene prediction (54,55,78). Each gene served as a sample in the machine learning procedure, either labelled as HDF or non-HDF, or was not used for training the classifiers.…”