PhaTYP: predicting the lifestyle for bacteriophages using BERT

Abstract

Bacteriophages (or phages), which infect bacteria, have two distinct lifestyles: virulent and temperate. Predicting the lifestyle of phages helps decipher their interactions with their bacterial hosts, aiding phages’ applications in fields such as phage therapy. Because experimental methods for annotating the lifestyle of phages cannot keep pace with the fast accumulation of sequenced phages, computational method for predicting phages’ lifestyles has become an attractive alternative. Despite some promising results, computational lifestyle prediction remains difficult because of the limited known annotations and the sheer amount of sequenced phage contigs assembled from metagenomic data. In particular, most of the existing tools cannot precisely predict phages’ lifestyles for short contigs. In this work, we develop PhaTYP (Phage TYPe prediction tool) to improve the accuracy of lifestyle prediction on short contigs. We design two different training tasks, self-supervised and fine-tuning tasks, to overcome lifestyle prediction difficulties. We rigorously tested and compared PhaTYP with four state-of-the-art methods: DeePhage, PHACTS, PhagePred and BACPHLIP. The experimental results show that PhaTYP outperforms all these methods and achieves more stable performance on short contigs. In addition, we demonstrated the utility of PhaTYP for analyzing the phage lifestyle on human neonates’ gut data. This application shows that PhaTYP is a useful means for studying phages in metagenomic data and helps extend our understanding of microbial communities.

Characterization of Klebsiella pneumoniae bacteriophages, KP1 and KP12, with deep learning-based structure prediction

“…In the present work, Klebsiella phages KP1 and KP12 were characterized and assigned as members of two distantly related Myoviridae clusters based on their TEMs and phylogenetic analyses. The lifestyle of the bacteriophages were additionally investigated by searching the presence of lysogeny-associated genes such as integrase and using several published prediction tools (e.g., PHACTS; McNair et al, 2012 ), PhagePred ( Song, 2020 ), BACHPHLIP ( Hockenberry and Wilke, 2021 ), and PhaTYP ( Shang et al, 2022 ; Supplementary Table S8 ). In case of KP1, there was no temperate phage marker genes found in the BLASTp analysis ( Supplementary Table S2 ), and there was a consistent prediction of KP1 as a lytic phage across the tools.…”

Leveraging transformers‐based language models in proteome bioinformatics

“…Many proteins have similar sequences or structures but perform different functions, while others have different sequences or structures but perform the same function. Previously, researchers applied NLP models to predict protein functions from gene ontology or textual terms such as learning representations [67] or predicting the lifestyle of bacteriophages [68]. Additionally, Cheng introduced BERTMHC, a more advanced MHCpeptide class II interaction prediction model that employs transformers and multiple instance learning [69].…”

Enhanced Bacterium–Phage Symbiosis in Attached Microbial Aggregates on a Membrane Surface Facing Elevated Hydraulic Stress

“…48 All identified viral genomes were taxonomically classified and clustered with viral genomes from the NCBI Viral RefSeq database by vConTACT2 (v0.9.19). 49 The lifestyle of phage contigs was identified by DeePhage (v1.0) 50 (virtual machine version), PhaTYP 51 (https://phage.ee.cityu.edu.hk/), and lysogenic markers detection by using hmmscan based on the list of lysogenic markers genes (transposase, integrase, excisionase, resolvase, and recombinase) from Pfam-A database 52 (Table S9). DeePhage (v1.0) was used to identify phage lifestyle with 0.5 as the cutoff of lysogeny, which was verified by the above PhaTYP online version.…”