bioRxiv Science⌕ Search

Biology subjects

Wiatrak, M.

Publications and source records attributed to Wiatrak, M..

2 recordsLinked to original sources

A contextualised protein language model reveals the functional syntax of bacterial evolution

Bacteria have evolved a vast diversity of functions and behaviours that are currently incompletely understood and poorly predicted from DNA sequence alone. To understand the syntax of bacterial evolution and discover genome-to-phenotype relationships, we curated over 1.3 million genomes spanning bacterial phylogenetic space, represented each as an ordered sequence of proteins, and used these sequences to train a transformer-based, contextualised protein language model, Bacformer. By pretraining on genome-wide evolutionary patterns, Bacformer captures the compositional and positional relationships of proteins and thereby provides a whole-genome framework for linking genomic organisation and content to measurable bacterial traits. We demonstrate the ability of Bacformer to accurately predict protein-protein interactions; uncover operon structure, which we validated experimentally; infer important phenotypic traits, including antimicrobial resistance, while revealing likely causal genes; and design template synthetic proteomes with desirable properties. Thus, Bacformer establishes a genomic foundation model that reveals the evolutionary rules governing bacterial gene organisation, function, and phenotype, opening a route to systematic whole-genome engineering.

microbiology↗

Sequence-based modelling of bacterial genomes enables accurate antibiotic resistance prediction

AbstractRapid detection of antibiotic-resistant bacteria and understanding the mecha- nisms underlying antimicrobial resistance (AMR) are major unsolved problems that pose significant threats to global public health. However, existing methods for predicting antibiotic resistance from genomic sequence data have had lim- ited success due to their inability to model epistatic effects and generalize to novel variants. Here, we present GeneBac, a deep learning method for predicting antibiotic resistance from DNA sequence through the integration of interactions between genes. We apply GeneBac to two distinct bacterial species and show that it can successfully predict the minimum inhibitory concentration (MIC) of multiple antibiotics. We use the WHO Mycobacterium tuberculosis mutation cat- alogue to demonstrate that GeneBac accurately predicts the effects of different variants, including novel variants that have not been observed during training. GeneBac is a modular framework which can be applied to a number of tasks including gene expression prediction, resistant gene identification and strain clus- tering. We leverage this modularity to transfer learn from the transcriptomic data to improve performance on the MIC prediction task.

microbiology↗