Search

Phages play critical roles in the survival and pathogenicity of their hosts, via lysogenic conversion factors, and in nutrient redistribution, via cell lysis. Analyses of phage- and viral-encoded genes in environmental samples provide insights into the physiological impact of viruses on microbial communities and human health. However, phage ORFs are extremely diverse of which over 70% of them are dissimilar to any genes with annotated functions in GenBank. Better identification of viruses would also aid in better detection and diagnosis of disease, in vaccine development, and generally in better understanding the physiological potential of any environment. In contrast to enzymes, viral structural protein function can be much more challenging to detect from sequence data because of low sequence conservation, few known conserved catalytic sites or sequence domains, and relatively limited experimental data. We have designed a method of predicting phage structural protein sequences that uses Artificial Neural Networks (ANNs). First, we trained ANNs to classify viral structural proteins using amino acid frequency; these correctly classify a large fraction of test cases with a high degree of specificity and sensitivity. Subsequently, we added estimates of protein isoelectric points as a feature to ANNs that classify specialized families of proteins, namely major capsid and tail proteins. As expected, these more specialized ANNs are more accurate than the structural ANNs. To experimentally validate the ANN predictions, several ORFs with no significant similarities to known sequences that are ANN-predicted structural proteins were examined by transmission electron microscopy. Some of these self-assembled into structures strongly resembling virion structures. Thus, our ANNs are new tools for identifying phage and potential prophage structural proteins that are difficult or impossible to detect by other bioinformatic analysis. The networks will be valuable when sequence is available but in vitro propagation of the phage may not be practical or possible., Author Summary Bacteriophages are extremely abundant and diverse biological entities. All phage particles are comprised of nucleic acids and structural proteins, with few other packaged proteins. Despite their simplicity and abundance, more than 70% of phage sequences in the viral Reference Sequence database encode proteins with unknown function based on FASTA annotations. As a result, the use of sequence similarity is often insufficient for detecting virus structural proteins among unknown viral sequences. Viral structural protein function is challenging to detect from sequence data because structural proteins possess few known conserved catalytic motifs and sequence domains. To address these issues we investigated the use of Artificial Neural Networks as an alternative means of predicting function. Here, we trained thousands of networks using the amino acid frequency of structural protein sequences and identified the optimal architectures with the highest accuracies. Some hypothetical protein sequences detected by our networks were expressed and visualized by TEM, and produced images that strongly resemble virion structures. Our results support the utility of our neural networks in predicting the functions of unknown viral sequences.

Crystallization of a maltose-binding protein MCL1 fusion has yielded a robust crystallography platform that generated the first apo MCL1 crystal structure, as well as five ligand-bound structures. The ability to obtain fragment-bound structures advances structure-based drug design efforts that, despite considerable effort, had previously been intractable by crystallography. In the ligand-independent crystal form we identify inhibitor binding modes not observed in earlier crystallographic systems. This MBP-MCL1 construct dramatically improves the structural understanding of well-validated MCL1 ligands, and will likely catalyze the structure-based optimization of high affinity MCL1 inhibitors.