BackgroundA major bottleneck in our understanding of the molecular underpinnings of life is the assignment of function to proteins. While molecular experiments provide the most reliable annotation of proteins, their relatively low throughput and restricted purview have led to an increasing role for computational function prediction. However, assessing methods for protein function prediction and tracking progress in the field remain challenging.ResultsWe conducted the second critical assessment of functional annotation (CAFA), a timed challenge to assess computational methods that automatically assign protein function. We evaluated 126 methods from 56 research groups for their ability to predict biological functions using Gene Ontology and gene-disease associations using Human Phenotype Ontology on a set of 3681 proteins from 18 species. CAFA2 featured expanded analysis compared with CAFA1, with regards to data set size, variety, and assessment metrics. To review progress in the field, the analysis compared the best methods from CAFA1 to those of CAFA2.ConclusionsThe top-performing methods in CAFA2 outperformed those from CAFA1. This increased accuracy can be attributed to a combination of the growing number of experimental annotations and improved methods for function prediction. The assessment also revealed that the definition of top-performing algorithms is ontology specific, that different performance metrics can be used to probe the nature of accurate predictions, and the relative diversity of predictions in the biological process and human phenotype ontologies. While there was methodological improvement between CAFA1 and CAFA2, the interpretation of results and usefulness of individual methods remain context-dependent.Electronic supplementary materialThe online version of this article (doi:10.1186/s13059-016-1037-6) contains supplementary material, which is available to authorized users.

BackgroundThe number of genomics and proteomics experiments is growing rapidly, producing an ever-increasing amount of data that are awaiting functional interpretation. A number of function prediction algorithms were developed and improved to enable fast and automatic function annotation. With the well-defined structure and manual curation, Gene Ontology (GO) is the most frequently used vocabulary for representing gene functions. To understand relationship and similarity between GO annotations of genes, it is important to have a convenient pipeline that quantifies and visualizes the GO function analyses in a systematic fashion.ResultsNaviGO is a web-based tool for interactive visualization, retrieval, and computation of functional similarity and associations of GO terms and genes. Similarity of GO terms and gene functions is quantified with six different scores including protein-protein interaction and context based association scores we have developed in our previous works. Interactive navigation of the GO function space provides intuitive and effective real-time visualization of functional groupings of GO terms and genes as well as statistical analysis of enriched functions.ConclusionsWe developed NaviGO, which visualizes and analyses functional similarity and associations of GO terms and genes. The NaviGO webserver is freely available at: http://kiharalab.org/web/navigo.

Motivation: Moonlighting proteins (MPs) show multiple cellular functions within a single polypeptide chain. To understand the overall landscape of their functional diversity, it is important to establish a computational method that can identify MPs on a genome scale. Previously, we have systematically characterized MPs using functional and omics-scale information. In this work, we develop a computational prediction model for automatic identification of MPs using a diverse range of protein association information. Results: We incorporated a diverse range of protein association information to extract characteristic features of MPs, which range from gene ontology (GO), protein-protein interactions, gene expression, phylogenetic profiles, genetic interactions and network-based graph properties to protein structural properties, i.e. intrinsically disordered regions in the protein chain. Then, we used machine learning classifiers using the broad feature space for predicting MPs. Because many known MPs lack some proteomic features, we developed an imputation technique to fill such missing features. Results on the control dataset show that MPs can be predicted with over 98% accuracy when GO terms are available. Furthermore, using only the omics-based features the method can still identify MPs with over 75% accuracy. Last, we applied the method on three genomes: Saccharomyces cerevisiae, Caenorhabditis elegans and Homo sapiens, and found that about 2-10% of proteins in the genomes are potential MPs. Availability and Implementation: Code available at http://kiharalab.org/MPprediction

BackgroundAdvancements in function prediction algorithms are enabling large scale computational annotation for newly sequenced genomes. With the increase in the number of functionally well characterized proteins it has been observed that there are many proteins involved in more than one function. These proteins characterized as moonlighting proteins show varied functional behavior depending on the cell type, localization in the cell, oligomerization, multiple binding sites, etc. The functional diversity shown by moonlighting proteins may have significant impact on the traditional sequence based function prediction methods. Here we investigate how well diverse functions of moonlighting proteins can be predicted by some existing function prediction methods.ResultsWe have analyzed the performances of three major sequence based function prediction methods, PSI-BLAST, the Protein Function Prediction (PFP), and the Extended Similarity Group (ESG) on predicting diverse functions of moonlighting proteins. In predicting discrete functions of a set of 19 experimentally identified moonlighting proteins, PFP showed overall highest recall among the three methods. Although ESG showed the highest precision, its recall was lower than PSI-BLAST. Recall by PSI-BLAST greatly improved when BLOSUM45 was used instead of BLOSUM62.ConclusionWe have analyzed the performances of PFP, ESG, and PSI-BLAST in predicting the functional diversity of moonlighting proteins. PFP shows overall better performance in predicting diverse moonlighting functions as compared with PSI-BLAST and ESG. Recall by PSI-BLAST greatly improved when BLOSUM45 was used. This analysis indicates that considering weakly similar sequences in prediction enhances the performance of sequence based AFP methods in predicting functional diversity of moonlighting proteins. The current study will also motivate development of novel computational frameworks for automatic identification of such proteins.

MotivationMoonlighting proteins (MPs) are an important class of proteins that perform more than one independent cellular function. MPs are gaining more attention in recent years as they are found to play important roles in various systems including disease developments. MPs also have a significant impact in computational function prediction and annotation in databases. Currently MPs are not labeled as such in biological databases even in cases where multiple distinct functions are known for the proteins. In this work, we propose a novel method named DextMP, which predicts whether a protein is a MP or not based on its textual features extracted from scientific literature and the UniProt database.ResultsDextMP extracts three categories of textual information for a protein: titles, abstracts from literature, and function description in UniProt. Three language models were applied and compared: a state-of-the-art deep unsupervised learning algorithm along with two other language models of different types, Term Frequency-Inverse Document Frequency in the bag-of-words and Latent Dirichlet Allocation in the topic modeling category. Cross-validation results on a dataset of known MPs and non-MPs showed that DextMP successfully predicted MPs with over 91% accuracy with significant improvement over existing MP prediction methods. Lastly, we ran DextMP with the best performing language models and text-based feature combinations on three genomes, human, yeast and Xenopus laevis, and found that about 2.5–35% of the proteomes are potential MPs.Availability and ImplementationCode available at http://kiharalab.org/DextMP.