15 February 2024 in PLOS Genetics
Elucidating gene function is a major goal in biology, especially among non-model organisms. However, doing so is complicated by the fact that molecular conservation does not always mirror functional conservation, and that complex relationships among genes are responsible for encoding pathways and higher-order biological processes. Co-expression, a promising approach for predicting gene function, relies on the general principal that genes with similar expression patterns across multiple conditions will likely be involved in the same biological process. For Cryptococcus neoformans, a prevalent human fungal pathogen greatly diverged from model yeasts, approximately 60% of the predicted genes in the genome lack functional annotations. Here, we leveraged a large amount of publicly available transcriptomic data to generate a C. neoformans Co-Expression Network (CryptoCEN), successfully recapitulating known protein networks, predicting gene function, and enabling insights into the principles influencing co-expression. With 100% predictive accuracy, we used CryptoCEN to identify 13 new DNA damage response genes, underscoring the utility of guilt-by-association for determining gene function. Overall, co-expression is a powerful tool for uncovering gene function, and decreases the experimental tests needed to identify functions for currently under-annotated genes.
18 November 2021 in PLOS ONE
The diversity of bacteriophages is likely unparalleled in the biome due to the immense variety of hosts and the multitude of viruses that infect them. Recent efforts have led to description at the genomic level of numerous bacteriophages that infect the Actinobacteria, but relatively little is known about those infecting other prokaryotic phyla, such as the purple non-sulfur photosynthetic α-proteobacterium Rhodobacter capsulatus. This species is a common inhabitant of freshwater ecosystems and has been an important model system for the study of photosynthesis. Additionally, it is notable for its utilization of a unique form of horizontal gene transfer via a bacteriophage-like element known as the gene transfer agent (RcGTA). Only three bacteriophages of R. capsulatus had been sequenced prior to this report. Isolation and characterization at the genomic level of 26 new bacteriophages infecting this host advances the understanding of bacteriophage diversity and the origins of RcGTA. These newly discovered isolates can be grouped along with three that were previously sequenced to form six clusters with four remaining as single representatives. These bacteriophages share genes with RcGTA that seem to be related to host recognition. One isolate was found to cause lysis of a marine bacterium when exposed to high-titer lysate. Although some clusters are more highly represented in the sequenced genomes, it is evident that many more bacteriophage types that infect R. capsulatus are likely to be found in the future.
30 March 2026 in Nature Communications
A central tenet of biology is that protein structure mediates the sequence-function relationship. Recently, there has been excitement about the promise of advances in protein structure modeling to generate hypotheses about sequence-structure-function relationships. Here, we leverage structural similarity to identify rapidly evolving proteasome assembly chaperones and characterize their function in Candidozyma (Candida) auris. Despite extensive sequence divergence, we demonstrate conservation of function, corroborating that specific folds, and not sequences, are required for function. This theoretical premise suggests that protein structures with certain properties should be functionally interchangeable, even if they were not products of a common evolutionary history. To reduce this theory to practice, we performed structure-informed protein design, exploring sequence space that is not accessible via stepwise evolution, and mutated more than 40 residues in the Poc4 proteasome assembly chaperone to demonstrate that artificial proteins can rescue complex biological processes in the context of the whole cell. This sequence-structure-function relationship expands our ability to use structure to identify deep evolutionary relationships between proteins and generate hypotheses about gene function in non-model organisms. Overall, this helps to define and understand functional constraints on protein evolution, with important implications for both future protein design and retrospective function prediction.