Biocatalysis is a promising approach to sustainably synthesize pharmaceuticals, complex natural products, and commodity chemicals at scale. However, the adoption of biocatalysis is limited by our ability to select enzymes that will catalyze their natural chemical transformation on non-natural substrates. While machine learning and in silico directed evolution are well-posed for this predictive modeling challenge, efforts to date have primarily aimed to increase activity against a single known substrate, rather than to identify enzymes capable of acting on new substrates of interest. To address this need, we curate 6 different high-quality enzyme family screens from the literature that each measure multiple enzymes against multiple substrates. We compare machine learning-based compound-protein interaction (CPI) modeling approaches from the literature used for predicting drug-target interactions. Surprisingly, comparing these interaction-based models against collections of independent (single task) enzyme-only or substrate-only models reveals that current CPI approaches are incapable of learning interactions between compounds and proteins in the current family level data regime. We further validate this observation by demonstrating that our no-interaction baseline can outperform CPI-based models from the literature used to guide the discovery of kinase inhibitors. Given the high performance of non-interaction based models, we introduce a new structure-based strategy for pooling residue representations across a protein sequence. Altogether, this work motivates a principled path forward in order to build and evaluate meaningful predictive models for biocatalysis and other drug discovery applications.
Bacteria and archaea have evolved with the ability to fix atmospheric dinitrogen in the form of ammonia, catalyzed by the nitrogenase enzyme complex which comprises three structural genes nifK, nifD and nifH. The nifK and nifD encodes for the beta and alpha subunits, respectively, of component 1, while nifH encodes for component 2 of nitrogenase. Phylogeny based on nifDHK have indicated that Cyanobacteria is closer to Proteobacteria alpha and gamma but not supported by the tree based on 16SrRNA. The evolutionary ancestor for the different trees was also different. The GC1 and GC2% analysis showed more consistency than GC3% which appeared to below for Firmicutes, Cyanobacteria and Euarchaeota while highest in Proteobacteria beta and clearly showed the proportional effect on the codon usage with a few exceptions. Few genes from Firmicutes, Euryarchaeota, Proteobacteria alpha and delta were found under mutational pressure. These nif genes with low and high GC3% from different classes of organisms showed similar expected number of codons. Distribution of the genes and codons, based on codon usage demonstrated opposite pattern for different orientation of mirror plane when compared with each other. Overall our results provide a comprehensive analysis on the evolutionary relationship of the three structural nif genes, nifK, nifD and nifH, respectively, in the context of codon usage bias, GC content relationship and amino acid composition of the encoded proteins and exploration of crucial statistical method for the analysis of positive data with non-constant variance to identify the shape factors of codon adaptation index.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.