EnzymeSifter: a tool for discovery of industrial enzymes from metagenomes

Abstract

Abstract Summary Metagenomes contain vast amounts of sequences, complicating the process of identifying candidate enzymes for industrial applications. Industrial applications require evaluating multiple biochemical properties simultaneously, including solubility, thermal stability, and pH. While separate predictors for each property exist, a score that combines multiple predicted values will be more descriptive than those generated individually by distinct tools. We present EnzymeSifter, a tool that automates enzyme discovery from vast metagenomes and enables multi-property predictions. It identifies the best performing enzymes using a computed composite score of all predicted values and generates a phylogenetic tree to select the top candidate from each clade—ensuring diversity and even sampling of sequence space. It acts as a sieve that filters according to the user inputs and keeps the most promising non-redundant enzymes for experimental validation. Availability and implementation EnzymeSifter is freely available and released under an MIT licence. EnzymeSifter source code is available at https://github.com/Bashton-Lab/EnzymeSifter