Search

The assignment of function to proteins at a large scale is essential for understanding the molecular mechanism of life. However, only a very small percentage of the more than 179 million proteins in UniProtKB have Gene Ontology (GO) annotations supported by experimental evidence. In this paper, we proposed an integrated deep-learning-based classification model, named SDN2GO, to predict protein functions. SDN2GO applies convolutional neural networks to learn and extract features from sequences, protein domains, and known PPI networks, and then utilizes a weight classifier to integrate these features and achieve accurate predictions of GO terms. We constructed the training set and the independent test set according to the time-delayed principle of the Critical Assessment of Function Annotation (CAFA) and compared it with two highly competitive methods and the classic BLAST method on the independent test set. The results show that our method outperforms others on each sub-ontology of GO. We also investigated the performance of using protein domain information. We learned from the Natural Language Processing (NLP) to process domain information and pre-trained a deep learning sub-model to extract the comprehensive features of domains. The experimental results demonstrate that the domain features we obtained are much improved the performance of our model. Our deep learning models together with the data pre-processing scripts are publicly available as an open source software at https://github.com/Charrick/SDN2GO.

Background Functional annotation of bacterial genomes is an obligatory and crucially important step of information processing from the genome sequences into cellular mechanisms. However, there is a lack of computational methods to evaluate the quality of functional assignments. Results We developed a genome-scale model that assigns Bayesian probability to each gene utilizing a known property of functional similarity between neighboring genes in bacteria. Conclusions Our model clearly distinguished true annotation from random annotation with Bayesian annotation probability >0.95. Our model will provide a useful guide to quantitatively evaluate functional annotation methods and to detect gene sets with reliable annotations. Electronic supplementary material The online version of this article (doi:10.1186/s12859-017-1761-2) contains supplementary material, which is available to authorized users.

As the biological datasets accumulate rapidly, computational methods designed to automate protein function prediction are critically needed. The problem of protein function prediction can be considered as a multi-label classification problem resulting in protein functional annotations. Nevertheless, biologists prefer to discover the correlations between protein attributes and functions. We introduce a multi-label supervised topic model into protein function prediction and investigate the advantages of this approach. This topic model can not only work out the function probability distributions over protein instances effectively, but also directly provide the words probability distributions over functions. To the best of our knowledge, this is the first effort to apply a multi-label supervised topic model to the protein function prediction. In this paper, we model a protein as a document and a function label as a topic. First, a set of protein sequences is formalized into a bag of words. Then, we perform inference and estimate the model parameters to predict protein functions. Experimental results on yeast and human datasets demonstrate the effectiveness of this multi-label supervised topic model on protein function prediction. Meanwhile, the experiments also show that this multi-label supervised topic model delivers superior results over the compared algorithms. In summary, the method discussed in this paper provides a new efficient approach to protein function prediction and reveals more information about functions.

In 2018, we reported a hybrid pipeline that predicts protein structures with I-TASSER and function with COFACTOR. I-TASSER/COFACTOR achieved Gene Ontology (GO) high prediction accuracies of Fmax = 0.69 and 0.57 for molecular function (MF) and biological process (BP), respectively, on 100 comprehensively annotated proteins. Now we report blinded analyses of newly annotated proteins in the critical assessment of function annotation (CAFA) three function prediction challenge and in neXtProt. For CAFA3 results released in May 2019, our predictions on 267 and 912 human proteins with newly annotated MF and BP terms achieved Fmax = 0.50 and 0.42, respectively, on "No Knowledge" proteins, and 0.51 and 0.74, respectively, on "Limited Knowledge" proteins. While COFACTOR consistently outperforms simple homology-based analysis, its accuracy still depends on template availability. Meanwhile, in neXtProt 2019-01, 25 proteins acquired new function annotation through literature curation at UniProt/Swiss-Prot. Before the release of these curated results, we submitted to neXtProt blinded predictions of free-text function annotation based on predicted GO terms. For 10 of the 25, a good match of free-text or GO term annotation was obtained. These blind tests represent rigorous assessments of I-TASSER/COFACTOR. neXtProt now provides links to precomputed I-TASSER/COFACTOR predictions for proteins without function annotation to facilitate experimental planning on "dark proteins".

We created a new high‐coverage, robust, and reproducible functional annotation of maize protein‐coding genes based on Gene Ontology (GO) term assignments. Whereas the existing Phytozome and Gramene maize GO annotation sets only cover 41% and 56% of maize protein‐coding genes, respectively, this study provides annotations for 100% of the genes. We also compared the quality of our newly derived annotations with the existing Gramene and Phytozome functional annotation sets by comparing all three to a manually annotated gold standard set of 1,619 genes where annotations were primarily inferred from direct assay or mutant phenotype. Evaluations based on the gold standard indicate that our new annotation set is measurably more accurate than those from Phytozome and Gramene. To derive this new high‐coverage, high‐confidence annotation set, we used sequence similarity and protein domain presence methods as well as mixed‐method pipelines that were developed for the Critical Assessment of Function Annotation (CAFA) challenge. Our project to improve maize annotations is called maize‐GAMER (GO Annotation Method, Evaluation, and Review), and the newly derived annotations are accessible via MaizeGDB (http://download.maizegdb.org/maize-GAMER) and CyVerse (B73 RefGen_v3 5b+ at doi.org/10.7946/P2S62P and B73 RefGen_v4 Zm00001d.2 at doi.org/10.7946/P2M925).

As a result of the genome sequencing and structural genomics initiatives, we have a wealth of protein sequence and structural data. However, only about 1% of these proteins have experimental functional annotations. As a result, computational approaches that can predict protein functions are essential in bridging this widening annotation gap. This article reviews the current approaches of protein function prediction using structure and sequence based classification of protein domain family resources with a special focus on functional families in the CATH-Gene3D resource.

Protein function prediction from sequence using the Gene Ontology (GO) classification is useful in many biological problems. It has recently attracted increasing interest, thanks in part to the Critical Assessment of Function Annotation (CAFA) challenge. In this paper, we introduce Guilty by Association on STRING (GAS), a tool to predict protein function exploiting protein-protein interaction networks without sequence similarity. The assumption is that whenever a protein interacts with other proteins, it is part of the same biological process and located in the same cellular compartment. GAS retrieves interaction partners of a query protein from the STRING database and measures enrichment of the associated functional annotations to generate a sorted list of putative functions. A performance evaluation based on CAFA metrics and a fair comparison with optimized BLAST similarity searches is provided. The consensus of GAS and BLAST is shown to improve overall performance. The PPI approach is shown to outperform similarity searches for biological process and cellular compartment GO predictions. Moreover, an analysis of the best practices to exploit protein-protein interaction networks is also provided.

Protein coding gene annotation errors in prokaryotic genomes are accumulating continually in bioinformatics databases, while the update rate of genome annotation can not keep up with the explosive increasing genome sequences in most cases. Hence it is critical to manually rectify the genome annotation errors. In this paper, a hybrid strategy by combing the gene ab initio predicting programs and the over annotated gene re-annotation programs is proposed for re-annotation of the protein coding genes in prokaryotic genomes. Based on this strategy, the protein coding genes in Geobacter sulfurreducens PCA is comprehensively re-annotated. As a consequence, 16 hypothetical genes are annotated as non-coding sequences and 104 missing genes are retrieved as protein coding genes. Subsequent function analysis and sequences analysis show that the predicting results are much reliable and robust. Further application to other genomes show that this work can provide alternative tools for later post-process of prokaryotic genome annotations.

With the development of next generation sequencing techniques, it is fast and cheap to determine protein sequences but relatively slow and expensive to extract useful information from protein sequences because of limitations of traditional biological experimental techniques. Protein function prediction has been a long standing challenge to fill the gap between the huge amount of protein sequences and the known function. In this paper, we propose a novel method to convert the protein function problem into a language translation problem by the new proposed protein sequence language "ProLan" to the protein function language "GOLan", and build a neural machine translation model based on recurrent neural networks to translate "ProLan" language to "GOLan" language. We blindly tested our method by attending the latest third Critical Assessment of Function Annotation (CAFA 3) in 2016, and also evaluate the performance of our methods on selected proteins whose function was released after CAFA competition. The good performance on the training and testing datasets demonstrates that our new proposed method is a promising direction for protein function prediction. In summary, we first time propose a method which converts the protein function prediction problem to a language translation problem and applies a neural machine translation model for protein function prediction.

The Human Genome Project and advances in DNA sequencing technologies have revolutionized the identification of genetic disorders through the use of clinical exome sequencing. However, in a considerable number of patients, the genetic basis remains unclear. As clinicians begin to consider whole-genome sequencing, an understanding of the processes and tools involved and the factors to consider in the annotation of the structure and function of genomic elements that might influence variant identification is crucial. Here, we discuss and illustrate the strengths and weaknesses of approaches for the annotation and classification of important elements of protein-coding genes, other genomic elements such as pseudogenes and the non-coding genome, comparative-genomic approaches for inferring gene function, and new technologies for aiding genome annotation, as a practical guide for clinicians when considering pathogenic sequence variation. Complete and accurate annotation of structure and function of genome features has the potential to reduce both false-negative (from missing annotation) and false-positive (from incorrect annotation) errors in causal variant identification in exome and genome sequences. Re-analysis of unsolved cases will be necessary as newer technology improves genome annotation, potentially improving the rate of diagnosis.

Ontologies have proven to be useful for capturing and organizing knowledge as a hierarchical set of terms and their relationships. However, curating gene ontology data by hand requires specialized knowledge of certain field, which is inefficient. Thus inferring gene ontology from the exponentially increased biological data is getting hot. Based on the Library of Integrated Network-Based Cellular Signatures (LINCS) data we came up with the hypothesis that genes participate in analogous biological processes might affect cells accordantly. By assessing cellular response after genes were knock out we built a similarity matrix with the Gene Set Enrichment Analysis (GSEA) and clustered the genes with affinity propagation algorithm. Next we mapped the cluster result to gene ontology biological process data for annotation and enrichment analysis, which confirmed our hypothesis and made it possible to predict biological processes for unannotated genes from cellular response data after genes are knock out for the first time. We further validated the rationality from the gene ontology molecular function data.1

Elucidating biological function of proteins is a fundamental problem in molecular biology and bioinformatics. Conventionally, protein function is annotated based on homology using sequence similarity search tools such as BLAST and FASTA. These methods perform well when obvious homologs exist for a query sequence; however, they will not provide any functional information otherwise. As a result, the functions of many genes in newly sequenced genomes are left unknown, which await functional interpretation. Here, we introduce two webservers for function prediction methods, which effectively use distantly related sequences to improve function annotation coverage and accuracy: Protein Function Prediction (PFP) and Extended Similarity Group (ESG). These two methods have been tested extensively in various benchmark studies and ranked among the top in community-based assessments for computational function annotation, including Critical Assessment of Function Annotation (CAFA) in 2010-2011 (CAFA1) and 2013-2014 (CAFA2). Both servers are equipped with user-friendly visualizations of predicted GO terms, which provide intuitive illustrations of relationships of predicted GO terms. In addition to PFP and ESG, we also introduce NaviGO, a server for the interactive analysis of GO annotations of proteins. All the servers are available at http://kiharalab.org/software.php .

Microbes play important roles in almost every aspect of life, including human health and diseases. Facilitated by the rapid development of sequencing technologies, metagenomics research has accelerated the accumulation of genomic sequences of microbial species that had been inaccessible before. Analysis of the metagenomic sequencing data can reveal not only the species but also the functional composition of microbial communities. Here, we report a pipeline for functional annotation of metagenomic datasets. The pipeline is built from several programs that we have developed for metagenomic sequence analysis including a protein-coding gene predictor for short reads (or contigs) and a fast similarity search tool. Given a metagenomic dataset, the pipeline reports putative protein-coding genes (or gene fragments) and functional annotations of the genes in Gene Ontology (GO) terms and Enzyme Commission (EC) numbers, and potential metabolic pathways that are likely encoded by the metagenome. Fun4Me is available for download at https://sourceforge.net/projects/fun4me .

We outline a set of strategies to infer protein function from structure. The overall approach depends on extensive use of homology modeling, the exploitation of a wide range of global and local geometric relationships between protein structures and the use of machine learning techniques. The combination of modeling with broad searches of protein structure space defines a “structural BLAST” approach to infer function with high genomic coverage. Applications are described to the prediction of protein–protein and protein–ligand interactions. In the context of protein–protein interactions, our structure-based prediction algorithm, PrePPI, has comparable accuracy to high-throughput experiments. An essential feature of PrePPI involves the use of Bayesian methods to combine structure-derived information with non-structural evidence (e.g. co-expression) to assign a likelihood for each predicted interaction. This, combined with a structural BLAST approach significantly expands the range of applications of protein structure in the annotation of protein function, including systems level biological applications where it has previously played little role.

We previously reported a paradigm for large-scale phylogenomic analysis of gene families that takes advantage of the large corpus of experimentally supported Gene Ontology (GO) annotations. This 'GO Phylogenetic Annotation' approach integrates GO annotations from evolutionarily related genes across ∼100 different organisms in the context of a gene family tree, in which curators build an explicit model of the evolution of gene functions. GO Phylogenetic Annotation models the gain and loss of functions in a gene family tree, which is used to infer the functions of uncharacterized (or incompletely characterized) gene products, even for human proteins that are relatively well studied. Here, we report our results from applying this paradigm to two well-characterized cellular processes, apoptosis and autophagy. This revealed several important observations with respect to GO annotations and how they can be used for function inference. Notably, we applied only a small fraction of the experimentally supported GO annotations to infer function in other family members. The majority of other annotations describe indirect effects, phenotypes or results from high throughput experiments. In addition, we show here how feedback from phylogenetic annotation leads to significant improvements in the PANTHER trees, the GO annotations and GO itself. Thus GO phylogenetic annotation both increases the quantity and improves the accuracy of the GO annotations provided to the research community. We expect these phylogenetically based annotations to be of broad use in gene enrichment analysis as well as other applications of GO annotations.Database URL: http://amigo.geneontology.org/amigo.

The annotation of protein provides a considerable knowledge for the biologists in order to understand life at the molecular level. The computational annotation of protein function has therefore emerged as an important alternative given that the biological experiments are extremely laborious. A number of methods have been developed to computationally annotate proteins using standardized nomenclatures such as Gene Ontology. These methods are based on various independency assumptions for modeling the annotation problem. However, the recent network analysis reveals that the same protein with different interactions may perform different functions. In this paper, we take into account the topology of the protein-protein interaction network in order to propose a new representation of functions' ontology. We use the Bayesian network in order to model and to alter the structure of this ontology so as to create the new context specific ontology. We use this newly proposed structure for predicting the functions of the unlabeled proteins. We evaluate our method, called Context-Specific Ontology by the use of the Bayesian Network (ConSOn-BN), on the Saccharomyces cerevisiae protein-protein interaction network and we find that ConSOn-BN has enhanced results as compared to some known methods.

Functional characterization of genes and their protein products is essential to biological and clinical research. Yet, there is still no reliable way of assigning functional annotations to proteins in a high-throughput manner. In this article, the authors provide an introduction to the task of automated protein function prediction. They discuss about the motivation for automated protein function prediction, the challenges faced in this task, as well as some approaches that are currently available. In particular, they take a closer look at methods that use protein-protein interaction for protein function prediction, elaborating on their underlying techniques and assumptions, as well as their strengths and limitations.

Over the past decade there has been a growing acknowledgement that a large proportion of proteins within most proteomes contain disordered regions. Disordered regions are segments of the protein chain which do not adopt a stable structure. Recognition of disordered regions in a protein is of great importance for protein structure prediction, protein structure determination and function annotation as these regions have a close relationship with protein expression and functionality. As a result, a great many protein disorder prediction methods have been developed so far. Here, we present an overview of current protein disorder prediction methods including an analysis of their advantages and shortcomings. In order to help users to select alternative tools under different circumstances, we also evaluate 23 disorder predictors on the benchmark data of the most recent round of the Critical Assessment of protein Structure Prediction (CASP) and assess their accuracy using several complementary measures.

Pharmaceutical researchers must evaluate vast numbers of protein sequences and formulate innovative strategies for identifying valid targets and discovering leads against them as a way of accelerating drug discovery. The ever increasing number and diversity of novel protein sequences identified by genomic sequencing projects and the success of worldwide structural genomics initiatives have spurred great interest and impetus in the development of methods for accurate, computationally empowered protein function prediction and active site identification. Previously, in the absence of direct experimental evidence, homology-based protein function annotation remained the gold-standard for in silico analysis and prediction of protein function. However, with the continued exponential expansion of sequence databases, this approach is not always applicable, as fewer query protein sequences demonstrate significant homology to protein gene products of known function. As a result, several non-homology based methods for protein function prediction that are based on sequence features, structure, evolution, biochemical and genetic knowledge have emerged. Herein, we review current bioinformatic programs and approaches for protein function prediction/annotation and discuss their integration into drug discovery initiatives. The development of such methods to annotate protein functional sites and their application to large protein functional families is crucial to successfully utilizing the vast amounts of genomic sequence information available to drug discovery and development processes.

A key role for ontologies in bioinformatics is their use as a standardized, structured terminology, particularly to annotate the genes in a genome with functional and other properties. Since the output of many genome-scale experiments results in gene sets it is natural to ask if they share a common function. A standard approach is to apply a statistical test for overrepresentation of functional annotation, often within the gene ontology. In this article we propose an alternative to the standard approach that avoids problems in overrepresentation analysis due to statistical dependencies between ontology categories. We apply methods of feature construction and selection to preprocess gene ontology terms used for the annotation of gene sets and incorporate these features as input to a standard supervised machine-learning algorithm. Our approach is shown to allow the straightforward use of an ontology in the context of data sourced from multiple experiments to learn classifiers predicting gene function as part of a cellular response to environmental stress.

We present an outline of the Critical Assessment of Protein Structure Prediction (CASP) infrastructure implemented at the University of California, Davis, Protein Structure Prediction Center. The infrastructure supports selection and validation of prediction targets, collection of predictions, standard evaluation of submitted predictions, and presentation of results. The Center also supports information exchange relating to CASP experiments and structure prediction in general. Technical aspects of conducting the CASP8 experiment and relevant statistics are also provided.

Motivation: Computational assignment of protein function may be the single most vital application of bioinformatics in the post-genome era. These assignments are made based on various protein features, where one is the presence of identifiable domains. The relationship between protein domain content and function is important to investigate, to understand how domain combinations encode complex functions. Results: Two different models are presented on how protein domain combinations yield specific functions: one rule-based and one probabilistic. We demonstrate how these are useful for Gene Ontology annotation transfer. The first is an intuitive generalization of the Pfam2GO mapping, and detects cases of strict functional implications of sets of domains. The second uses a probabilistic model to represent the relationship between domain content and annotation terms, and was found to be better suited for incomplete training sets. We implemented these models as predictors of Gene Ontology functional annotation terms. Both predictors were more accurate than conventional best BLAST-hit annotation transfer and more sensitive than a single-domain model on a large-scale dataset. We present a number of cases where combinations of Pfam-A protein domains predict functional terms that do not follow from the individual domains. Availability: Scripts and documentation are available for download at http://sonnhammer.sbc.su.se/multipfam2go_source_docs.tar Contact: Kristoffer.Forslund@sbc.su.se Supplementary information: Supplementary data are available at Bioinformatics online.

This unit describes the concepts and practical techniques for annotating genomic variants in the human genome to estimate their functional significance. With the rapid increase of available whole exome and whole genome sequencing information for human studies, annotation techniques have become progressively more important for highlighting and prioritizing nucleotide variants and their potential impact on genes and other genetic constructs. Here, we present an overview of different types of variant annotation approaches and elaborate on their foundations, assumptions, and the downstream consequences of their use. Computational approaches and tools to assign annotations and to identify variants are reviewed. Further, the general philosophy of assigning potential function to a genetic change within the biological contexts of a disease is discussed.

A well-annotated genome is necessary for the rapid identification of genes in the genome, gene mapping and functional analysis. The arrangement of any genome has typically taken many years, but with new sequencing methodologies, the time to produce genomes is greatly reduced; however, with this increased speed, there has been a deterioration in the quality of genomes, with many genomes being published as drafts. The newly published draft kiwifruit genome is a fantastic starting point for genomics in kiwifruit. However, as with all genomes, there is an opportunity to improve both the construction and annotation. In this chapter, we detail the state of the current annotation and propose a new initiative to generate better annotation of kiwifruit genes.

Increasing evidences indicated that function annotation of human genome in molecular level and phenotype level is very important for systematic analysis of genes. In this study, we presented a framework named Gene2Function to annotate Gene Reference into Functions (GeneRIFs), in which each functional description of GeneRIFs could be annotated by a text mining tool Open Biomedical Annotator (OBA), and each Entrez gene could be mapped to Human Genome Organisation Gene Nomenclature Committee (HGNC) gene symbol. After annotating all the records about human genes of GeneRIFs, 288,869 associations between 13,148 mRNAs and 7,182 terms, 9,496 associations between 948 microRNAs and 533 terms, and 901 associations between 139 long noncoding RNAs (lncRNAs) and 297 terms were obtained as a comprehensive annotation resource of human genome. High consistency of term frequency of individual gene (Pearson correlation = 0.6401,p=2.2e-16) and gene frequency of individual term (Pearson correlation = 0.1298,p=3.686e-14) in GeneRIFs and GOA shows our annotation resource is very reliable.

Function prediction of uncharacterized protein sequences generated by genome projects has emerged as an important focus for computational biology. We have categorized several approaches beyond traditional sequence similarity that utilize the overwhelmingly large amounts of available data for computational function prediction, including structure-, association (genomic context)-, interaction (cellular context)-, process (metabolic context)-, and proteomics-experiment-based methods. Because they incorporate structural and experimental data that is not used in sequence-based methods, they can provide additional accuracy and reliability to protein function prediction. Here, first we review the definition of protein function. Then the recent developments of these methods are introduced with special focus on the type of predictions that can be made. The need for further development of comprehensive systems biology techniques that can utilize the ever-increasing data presented by the genomics and proteomics communities is emphasized. For the readers' convenience, tables of useful online resources in each category are included. The role of computational scientists in the near future of biological research and the interplay between computational and experimental biology are also addressed.

I-TASSER is a hierarchical protocol for automated protein structure prediction and structure-based function annotation. Starting from the amino acid sequence of target proteins, I-TASSER first generates full-length atomic structural models from multiple threading alignments and iterative structural assembly simulations followed by atomic-level structure refinement. The biological functions of the protein, including ligand-binding sites, enzyme commission number, and gene ontology terms, are then inferred from known protein function databases based on sequence and structure profile comparisons. I-TASSER is freely available as both an on-line server and a stand-alone package. This unit describes how to use the I-TASSER protocol to generate structure and function prediction and how to interpret the prediction results, as well as alternative approaches for further improving the I-TASSER modeling quality for distant-homologous and multi-domain protein targets.

The function annotation process in computational biology has increasingly shifted from the traditional characterization of individual biochemical roles of protein molecules to the system-wide detection of entire metabolic pathways and genomic structures. The so-called genome-aware methods broaden misannotation inconsistencies in genome sequences beyond protein function assignments, encompassing phylogenetic anomalies and artifactual genomic regions. We outline three categories of error propagation in databases by providing striking examples – at various levels of appreciation by the community from traditional to emerging, thus raising awareness for future solutions.

Functional annotation is the process that assigns a biological functionality to a deoxyribonucleic acid (DNA) sequence. It requires searching in huge data sets for candidates, and inferring the most appropriate features based on the information found and expert knowledge. When humans perform most of these tasks, results are of a high quality, but there is a bottleneck in processing; when experts are largely replaced by automated tools, annotation is faster but of poorer quality. Combining the automatic annotation with expert systems (ESs) can enhance the quality of the annotation, while effectively reducing experts' workload. This paper presents INFAES, a rule-based ES developed for mimicking the human reasoning in the inference stage of the functional annotation. It integrates knowledge on Biology and heuristics about the use of Bioinformatics tools. Its development adopts state-of-the-art methodologies to facilitate the acquisition and integration of new knowledge. INFAES showed a high performance when compared to the systems developed for the first large-scale community-based critical assessment of protein function annotation (CAFA). © 2015 Elsevier B.V. All rights reserved., This work has been done in the context of the projects “Aquagenomics” (CSD2007-00002) funded by the Consolider-Ingenio 2010 program of the Spanish Ministry of Education and Science, and “Social Ambient Assisting Living - Methods (SociAAL)” (TIN2011-28335-C02-01) supported by the Spanish Ministry for Economy and Competitiveness. Also, we acknowledge support from the “Red Científico-Tecnológica en Ciencias de los Servicios” (TIN2011-15497-E) and the “Programa de Creación y Consolidación de Grupos de Investigación” (UCM-BSCH GR35/10-A).

Background The recently held Critical Assessment of Function Annotation challenge (CAFA2) required its participants to submit predictions for a large number of target proteins regardless of whether they have previous annotations or not. This is in contrast to the original CAFA challenge in which participants were asked to submit predictions for proteins with no existing annotations. The CAFA2 task is more realistic, in that it more closely mimics the accumulation of annotations over time. In this study we compare these tasks in terms of their difficulty, and determine whether cross-validation provides a good estimate of performance. Results The CAFA2 task is a combination of two subtasks: making predictions on annotated proteins and making predictions on previously unannotated proteins. In this study we analyze the performance of several function prediction methods in these two scenarios. Our results show that several methods (structured support vector machine, binary support vector machines and guilt-by-association methods) do not usually achieve the same level of accuracy on these two tasks as that achieved by cross-validation, and that predicting novel annotations for previously annotated proteins is a harder problem than predicting annotations for uncharacterized proteins. We also find that different methods have different performance characteristics in these tasks, and that cross-validation is not adequate at estimating performance and ranking methods. Conclusions These results have implications for the design of computational experiments in the area of automated function prediction and can provide useful insight for the understanding and design of future CAFA competitions. Electronic supplementary material The online version of this article (doi:10.1186/s13742-015-0082-5) contains supplementary material, which is available to authorized users.

The characterization of new genomes based on their protein sets has been revolutionized by new sequencing technologies, but biologists seeking to exploit new sequence information are often frustrated by the challenges associated with accurately assigning biological functions to newly identified proteins. Here, we highlight some of the challenges in functional inference from sequence similarity. Investigators can improve the accuracy of function prediction by (1) being conservative about the evolutionary distance to a protein of known function; (2) considering the ambiguous meaning of "functional similarity," and (3) being aware of the limitations of annotations in functional databases. Protein function prediction does not offer "one-size-fits-all" solutions. Prediction strategies work better when the idiosyncrasies of function and functional annotation are better understood.

Gene Ontology is one of the largest bioinformatics project that seeks to consolidate knowledge about genes through annotation of terms to three ontologies. In this work, we present a technique to find association relationships in the annotation terms for the Saccharomyces cerevisiae (SGD) genome. We first present a normalization algorithm to ensure that the annotation terms have a similar level of specificity. Association rule mining algorithms are used to find significant and non-trivial association rules in these normalized datasets. Metrics such as support, confidence, and lift can be used to evaluate the strength of found rules. We conducted experiments on the entire SGD annotation dataset and here we present the top 10 strongest rules for each of the three ontologies. We verify the found rules using evidence from the biomedical literature. The presented method has a number of advantages — it relies only on the structure of the gene ontology, has minimal memory and storage requirements, and can be easily scaled for large genomes, such as the human genome. There are many applications of this technique, such as predicting the GO annotations for new genes or those that have not been studied extensively.

Motivation: Systematically predicting gene (or protein) function based on molecular interaction networks has become an important tool in refining and enhancing the existing annotation catalogs, such as the Gene Ontology (GO) database. However, functional labels with only a few (, National Institute of General Medical Sciences (U.S.) (Grant 1U54GM114838)

In science, the relationship between methods and discovery is symbiotic. As we discover more, we are able to construct more precise and sensitive tools and methods that enable further discovery. With better lens crafting came microscopes, and with them the discovery of living cells. In the last 40 years, advances in molecular biology, statistics, and computer science have ushered in the field of bioinformatics and the genomic era. Computational scientists enjoy developing new methods, and the community encourages them to do so. Indeed, the editorial guidelines for PLOS Computational Biology require manuscripts to apply novel methods. However, it is often confusing to know which method to choose: which method is best? And, in this context, what does “best” mean? To help choose an appropriate method for a particular task, scientists often form community-based challenges for the unbiased evaluation of methods in a given field. These challenges help evaluate existing and novel methods, while helping to coalesce a community and leading to new ideas and collaborations. In computational biology, the first of these challenges was arguably the Critical Assessment of protein Structure Prediction, or CASP [1], whose goal is to evaluate methods for predicting three-dimensional protein structure from amino acid sequence. The first CASP meeting was held in December of 1994, following a “prediction period” where members of the community were presented with protein amino acid sequences and asked to predict their three dimensional structures. The sequences that were chosen had recently been solved by X-ray crystallography but had not been not published or released until after the predictions from the community were made. Since the first CASP, we have seen many successful challenges, including Critical Assessment of Function Annotation (CAFA) for protein function prediction [2], Critical Assessment of Genome Interpretation (CAGI) (for genome interpretation) [3], Critical Assessment of Massive (originally “Microarray”) Data Analysis (CAMDA) (for large-scale biological data) [4], BioCreative (for biomedical text mining) [5], the Assemblathon (for sequence assembly), and the NCI-DREAM Challenges (for various biomedical challenges), amongst others [6]. Computational challenges also help solve new problems. While the original CASP experiment was developed to evaluate existing methods applied to current problems, other communities often look at other areas for which there are no existing tools. These challenges have spread successfully to industry, and companies such as Innocentive [7] and X-Prize [8] offer large prizes for solving novel questions. Because these challenges are, on one hand, an exercise in community collaboration, and on the other, a competition, organizing a challenge is littered with difficulties and pitfalls. Having served as organizers, predictors, and assessors within several existing communities, we present ten rules we believe should be observed when organizing a computational methods challenge

Clustering is perhaps one of the most widely used tools for microarray data analysis. Proposed roles for genes of unknown function are inferred from clusters of genes similarity expressed across many biological conditions. However, whether function annotation by similarity metrics is reliable or not and to what extent the similarity in gene expression patterns is useful for annotation of gene functions, has not been evaluated. This paper made a comprehensive research on the correlation between the similarity of expression data and of gene functions using Gene Ontology. It has been found that although the similarity in expression patterns and the similarity in gene functions are significantly dependent on each other, this association is rather weak. In addition, among the three categories of Gene Ontology, the similarity of expression data is more useful for cellular component annotation than for biological process and molecular function. The results presented are interesting for the gene functions prediction research area.

The ability to predict the function of a protein, given its sequence and/or 3D structure, is an essential requirement for exploiting the wealth of data made available by genomics and structural genomics projects and is therefore raising increasing interest in the computational biology community. To foster developments in the area as well as to establish the state of the art of present methods, a function prediction category was tentatively introduced in the 6th edition of the Critical Assessment of Techniques for Protein Structure Prediction (CASP) worldwide experiment. The assessment of the performance of the methods was made difficult by at least two factors: (a) the experimentally determined function of the targets was not available at the time of assessment; (b) the experiment is run blindly, preventing verification of whether the convergence of different predictions towards the same functional annotation was due to the similarity of the methods or to a genuine signal detectable by different methodologies. In this work, we collected information about the methods used by the various predictors and revisited the results of the experiment by verifying how often and in which cases a convergent prediction was obtained by methods based on different rationale. We propose a method for classifying the type and redundancy of the methods. We also analyzed the cases in which a function for the target protein has become available. Our results show that predictions derived from a consensus of different methods can reach an accuracy as high as 80%. It follows that some of the predictions submitted to CASP6, once reanalyzed taking into account the type of converging methods, can provide very useful information to researchers interested in the function of the target proteins.

Automated function prediction (AFP) methods increasingly use knowledge discovery algorithms to map sequence, structure, literature, and/or pathway information about proteins whose functions are unknown into functional ontologies, typically (a portion of) the Gene Ontology (GO). While there are a growing number of methods within this paradigm, the general problem of assessing the accuracy of such prediction algorithms has not been seriously addressed. We present first an application for function prediction from protein sequences using the POSet Ontology Categorizer (POSOC) to produce new annotations by analyzing collections of GO nodes derived from annotations of protein BLAST neighborhoods. We then also present hierarchical precision and hierarchical recall as new evaluation metrics for assessing the accuracy of any predictions in hierarchical ontologies, and discuss results on a test set of protein sequences. We show that our method provides substantially improved hierarchical precision (measure of predictions made that are correct) when applied to the nearest BLAST neighbors of target proteins, as compared with simply imputing that neighborhood's annotations to the target. Moreover, when our method is applied to a broader BLAST neighborhood, hierarchical precision is enhanced even further. In all cases, such increased hierarchical precision performance is purchased at a modest expense of hierarchical recall (measure of all annotations that get predicted at all).

Overwhelmed with genomic data, biologists are facing the first big post-genomic question--what do all genes do? First, not only is the volume of pure sequence and structure data growing, but its diversity is growing as well, leading to a disproportionate growth in the number of uncharacterized gene products. Consequently, established methods of gene and protein annotation, such as homology-based transfer, are annotating less data and in many cases are amplifying existing erroneous annotation. Second, there is a need for a functional annotation which is standardized and machine readable so that function prediction programs could be incorporated into larger workflows. This is problematic due to the subjective and contextual definition of protein function. Third, there is a need to assess the quality of function predictors. Again, the subjectivity of the term 'function' and the various aspects of biological function make this a challenging effort. This article briefly outlines the history of automated protein function prediction and surveys the latest innovations in all three topics.

Motivation: Most approaches in predicting protein function from protein–protein interaction data utilize the observation that a protein often share functions with proteins that interacts with it (its level-1 neighbours). However, proteins that interact with the same proteins (i.e. level-2 neighbours) may also have a greater likelihood of sharing similar physical or biochemical characteristics. We speculate that functional similarity between a protein and its neighbours from the two different levels arise from two distinct forms of functional association, and a protein is likely to share functions with its level-1 and/or level-2 neighbours. We are interested in finding out how significant is functional association between level-2 neighbours and how they can be exploited for protein function prediction. Results: We made a statistical study on recent interaction data and observed that functional association between level-2 neighbours is clearly observable. A substantial number of proteins are observed to share functions with level-2 neighbours but not with level-1 neighbours. We develop an algorithm that predicts the functions of a protein in two steps: (1) assign a weight to each of its level-1 and level-2 neighbours by estimating its functional similarity with the protein using the local topology of the interaction network as well as the reliability of experimental sources and (2) scoring each function based on its weighted frequency in these neighbours. Using leave-one-out cross validation, we compare the performance of our method against that of several other existing approaches and show that our method performs relatively well. Contact: g0306417@nus.edu.sg

Automated annotation of high-throughput genome sequences is one of the earliest steps toward a comprehensive understanding of the dynamic behavior of living organisms. However, the step is often error-prone because of its underlying algorithms, which rely mainly on a simple similarity analysis, and lack of guidance from biological rules. We present herein a knowledge-based protein annotation algorithm. Our objectives are to reduce errors and to improve annotation confidences. This algorithm consists of two major components: a knowledge system, called “RuleMiner,” and a voting procedure. The knowledge system, which includes biological rules and functional profiles for each function, provides a platform for seamless integration of multiple sequence analysis tools and guidance for function annotation. The voting procedure, which relies on the knowledge system, is designed to make (possibly) unbiased judgments in functional assignments among complicated, sometimes conflicting, information. We have applied this algorithm to 10 prokaryotic bacterial genomes and observed a significant improvement in annotation confidences. We also discuss the current limitations of the algorithm and the potential for future improvement. Proteins 2005. © 2005 Wiley-Liss, Inc.

Motivation: Gene Ontology (GO) consortium provides structural description of protein function that is used as a common language for gene annotation in many organisms. Large-scale techniques have generated many valuable protein–protein interaction datasets that are useful for the study of protein function. Combining both GO and protein–protein interaction data allows the prediction of function for unknown proteins. Result: We apply a Markov random field method to the prediction of yeast protein function based on multiple protein–protein interaction datasets. We assign function to unknown proteins with a probability representing the confidence of this prediction. The functions are based on three general categories of cellular component, molecular function and biological process defined in GO. The yeast proteins are defined in the Saccharomyces Genome Database (SGD). The protein–protein interaction datasets are obtained from the Munich Information Center for Protein Sequences (MIPS), including physical interactions and genetic interactions. The efficiency of our prediction is measured by applying the leave-one-out validation procedure to a functional path matching scheme, which compares the prediction with the GO description of a protein's function from the abstract level to the detailed level along the GO structure. For biological process, the leave-one-out validation procedure shows 52% precision and recall of our method, much better than that of the simple guilty-by-association methods. Supplementary material: http://www.cmb.usc.edu/~msms/gomapping

Item does not contain fulltext The growing number of completely sequenced genomes adds new dimensions to the use of sequence analysis to predict protein function. Compared with the classical knowledge transfer from one protein to a similar sequence (homology-based function prediction), knowledge about the corresponding genes in other genomes (orthology-based function prediction) provides more specific information about the protein's function, while the analysis of the sequence in its genomic context (context-based function prediction) provides information about its functional context. Whereas homology-based methods predict the molecular function of a protein, genomic context methods predict the biological process in which it plays a role. These complementary approaches can be combined to elucidate complete functional networks and biochemical pathways from the genome sequence of an organism. Here we review recent advances in the field of genomic-context based methods of protein function prediction. Techniques are highlighted with examples, including an analysis that combines information from genomic-context with homology to predict a role of the RNase L inhibitor in the maturation of ribosomal RNA.

Assigning functions to novel proteins is one of the most important problems in the postgenomic era. Several approaches have been applied to this problem, including the analysis of gene expression patterns, phylogenetic profiles, protein fusions, and protein-protein interactions. In this paper, we develop a novel approach that employs the theory of Markov random fields to infer a protein's functions using protein-protein interaction data and the functional annotations of protein's interaction partners. For each function of interest and protein, we predict the probability that the protein has such function using Bayesian approaches. Unlike other available approaches for protein annotation in which a protein has or does not have a function of interest, we give a probability for having the function. This probability indicates how confident we are about the prediction. We employ our method to predict protein functions based on "biochemical function," "subcellular location," and "cellular role" for yeast proteins defined in the Yeast Proteome Database (YPD, www.incyte.com), using the protein-protein interaction data from the Munich Information Center for Protein Sequences (MIPS, mips.gsf.de). We show that our approach outperforms other available methods for function prediction based on protein interaction data. The supplementary data is available at www-hto.usc.edu/~msms/ProteinFunction.

Genomic sequencing is no longer a novelty, but gene function annotation remains a key challenge in modern biology. A variety of functional genomics experimental techniques are available, from classic methods such as affinity precipitation to advanced high-throughput techniques such as gene expression microarrays. In the future, more disparate methods will be developed, further increasing the need for integrated computational analysis of data generated by these studies. We address this problem with magic (Multisource Association of Genes by Integration of Clusters), a general framework that uses formal Bayesian reasoning to integrate heterogeneous types of high-throughput biological data (such as large-scale two-hybrid screens and multiple microarray analyses) for accurate gene function prediction. The system formally incorporates expert knowledge about relative accuracies of data sources to combine them within a normative framework. magic provides a belief level with its output that allows the user to vary the stringency of predictions. We applied magic to Saccharomyces cerevisiae genetic and physical interactions, microarray, and transcription factor binding sites data and assessed the biological relevance of gene groupings using Gene Ontology annotations produced by the Saccaromyces Genome Database. We found that by creating functional groupings based on heterogeneous data types, magic improved accuracy of the groupings compared with microarray analysis alone. We describe several of the biological gene groupings identified.

Motivation: The human genome project has led to the discovery of many human protein coding genes which were previously unknown. As a large fraction of these are functionally uncharacterized, it is of interest to develop methods for predicting their molecular function from sequence. Results: We have developed a method for prediction of protein function for a subset of classes from the Gene Ontology classification scheme. This subset includes several pharmaceutically interesting categories—transcription factors, receptors, ion channels, stress and immune response proteins, hormones and growth factors can all be predicted. Although the method relies on protein sequences as the sole input, it does not rely on sequence similarity, but instead on sequence derived protein features such as predicted post translational modifications (PTMs), protein sorting signals and physical/chemical properties calculated from the amino acid composition. This allows for prediction of the function for orphan proteins where no homologs can be found. Using this method we propose two novel receptors in the human genome, and further demonstrate chromosomal clustering of related proteins. Availability: Sequences can be submitted to the prediction server via a web interface at http://www.cbs.dtu.dk/services/ProtFun/ Contact: ljj@cbs.dtu.dk * To whom correspondence should be addressed.

The Critical Assessment of Function Annotation meeting was held July 14-15, 2011 at the Austria Conference Center in Vienna, Austria. There were 73 registered delegates at the meeting. We thank the DOE for this award. It helped us organize and support a scientific meeting AFP 2011 as a special interest group (SIG) meeting associated with the ISMB 2011 conference. The conference was held in Vienna, Austria, in July 2011. The AFP SIG was held on July 15-16, 2011 (immediately preceding the conference). The meeting consisted of two components, the first being a series of talks (invited and contributed) and discussion sections dedicated to protein function research, with an emphasis on the theory and practice of computational methods utilized in functional annotation. The second component provided a large-scale assessment of computational methods through participation in the Critical Assessment of Functional Annotation (CAFA).

Public sequence databases contain information on the sequence, structure and function of proteins. Genome sequencing projects have led to a rapid increase in protein sequence information, but reliable, experimentally verified, information on protein function lags a long way behind. To address this deficit, functional annotation in protein databases is often inferred by sequence similarity to homologous, annotated proteins, with the attendant possibility of error. Now, the functional annotation in these homologous proteins may itself have been acquired through sequence similarity to yet other proteins, and it is generally not possible to determine how the functional annotation of any given protein has been acquired. Thus the possibility of chains of misannotation arises, a process we term ‘error percolation’. With some simple assumptions, we develop a dynamical probabilistic model for these misannotation chains. By exploring the consequences of the model for annotation quality it is evident that this iterative approach leads to a systematic deterioration of database quality. Contact: WRG: wally.gilks@mrc-bsu.cam.ac.uk; BA and CAO: audit@ebi.ac.uk; ouzounis@ebi.ac.uk * To whom correspondence should be addressed. † Both these authors contributed equally to this work.

Background Functional annotation of novel proteins is one of the central problems in bioinformatics. With the ever-increasing development of genome sequencing technologies, more and more sequence information is becoming available to analyze and annotate. To achieve fast and automatic function annotation, many computational (automated) function prediction (AFP) methods have been developed. To objectively evaluate the performance of such methods on a large scale, community-wide assessment experiments have been conducted. The second round of the Critical Assessment of Function Annotation (CAFA) experiment was held in 2013–2014. Evaluation of participating groups was reported in a special interest group meeting at the Intelligent Systems in Molecular Biology (ISMB) conference in Boston in 2014. Our group participated in both CAFA1 and CAFA2 using multiple, in-house AFP methods. Here, we report benchmark results of our methods obtained in the course of preparation for CAFA2 prior to submitting function predictions for CAFA2 targets. Results For CAFA2, we updated the annotation databases used by our methods, protein function prediction (PFP) and extended similarity group (ESG), and benchmarked their function prediction performances using the original (older) and updated databases. Performance evaluation for PFP with different settings and ESG are discussed. We also developed two ensemble methods that combine function predictions from six independent, sequence-based AFP methods. We further analyzed the performances of our prediction methods by enriching the predictions with prior distribution of gene ontology (GO) terms. Examples of predictions by the ensemble methods are discussed. Conclusions Updating the annotation database was successful, improving the Fmax prediction accuracy score for both PFP and ESG. Adding the prior distribution of GO terms did not make much improvement. Both of the ensemble methods we developed improved the average Fmax score over all individual component methods except for ESG. Our benchmark results will not only complement the overall assessment that will be done by the CAFA organizers, but also help elucidate the predictive powers of sequence-based function prediction methods in general. Electronic supplementary material The online version of this article (doi:10.1186/s13742-015-0083-4) contains supplementary material, which is available to authorized users.

Data from large-scale genome projects, transcriptomics and proteomics experiments have provided scientists with a wealth of information establishing the basis for the investigation of cellular processes. To understand biological function beyond the single gene by the discovery and characterization of functional protein networks, bioinformatics analysis requires information about two additional attributes associated with the gene products: (i) high-level protein function prediction of experimentally uncharacterized proteins and (ii) systematic classification of protein function. This article describes the basic properties of protein classification systems and discusses examples of their implementation.