Predicting Gene Ontology Terms Research Articles

A large-scale assessment of sequence database search tools for homology-based protein function prediction.

Sequence database searches followed by homology-based function transfer form one of the oldest and most popular approaches for predicting protein functions, such as Gene Ontology (GO) terms. These searches are also a critical component in most state-of-the-art machine learning and deep learning-based protein function predictors. Although sequence search tools are the basis of homology-based protein function prediction, previous studies have scarcely explored how to select the optimal sequence search tools and configure their parameters to achieve the best function prediction. In this paper, we evaluate the effect of using different options from among popular search tools, as well as the impacts of search parameters, on protein function prediction. When predicting GO terms on a large benchmark dataset, we found that BLASTp and MMseqs2 consistently exceed the performance of other tools, including DIAMOND-one of the most popular tools for function prediction-under default search parameters. However, with the correct parameter settings, DIAMOND can perform comparably to BLASTp and MMseqs2 in function prediction. Additionally, we developed a new scoring function to derive GO prediction from homologous hits that consistently outperform previously proposed scoring functions. These findings enable the improvement of almost all protein function prediction algorithms with a few easily implementable changes in their sequence homolog-based component. This study emphasizes the critical role of search parameter settings in homology-based function transfer and should have an important contribution to the development of future protein function prediction algorithms.

FunPredCATH: An ensemble method for predicting protein function using CATH

MotivationThe growth of unannotated proteins in UniProt increases at a very high rate every year due to more efficient sequencing methods. However, the experimental annotation of proteins is a lengthy and expensive process. Using computational techniques to narrow the search can speed up the process by providing highly specific Gene Ontology (GO) terms. MethodologyWe propose an ensemble approach that combines three generic base predictors that predict Gene Ontology (BP, CC and MF) terms from sequences across different species. We train our models on UniProtGOA annotation data and use the CATH domain resources to identify the protein families. We then calculate a score based on the prevalence of individual GO terms in the functional families that is then used as an indicator of confidence when assigning the GO term to an uncharacterised protein. MethodsIn the ensemble, we use a statistics-based method that scores the occurrence of GO terms in a CATH FunFam against a background set of proteins annotated by the same GO term. We also developed a set-based method that uses Set Intersection and Set Union to score the occurrence of GO terms within the same CATH FunFam. Finally, we also use FunFams-Plus, a predictor method developed by the Orengo Group at UCL to predict GO terms for uncharacterised proteins in the CAFA3 challenge. EvaluationWe evaluated the methods against the CAFA3 benchmark and DomFun. We used the Precision, Recall and Fmax metrics and the benchmark datasets that are used in CAFA3 to evaluate our models and compare them to the CAFA3 results. Our results show that FunPredCATH compares well with top CAFA methods in the different ontologies and benchmarks. ContributionsFunPredCATH compares well with other prediction methods on CAFA3, and the ensemble approach outperforms the base methods. We show that non-IEA models obtain higher Fmax scores than the IEA counterparts, while the models including IEA annotations have higher coverage at the expense of a lower Fmax score.

SANA: cross-species prediction of Gene Ontology GO annotations via topological network alignment

Topological network alignment aims to align two networks node-wise in order to maximize the observed common connection (edge) topology between them. The topological alignment of two protein–protein interaction (PPI) networks should thus expose protein pairs with similar interaction partners allowing, for example, the prediction of common Gene Ontology (GO) terms. Unfortunately, no network alignment algorithm based on topology alone has been able to achieve this aim, though those that include sequence similarity have seen some success. We argue that this failure of topology alone is due to the sparsity and incompleteness of the PPI network data of almost all species, which provides the network topology with a small signal-to-noise ratio that is effectively swamped when sequence information is added to the mix. Here we show that the weak signal can be detected using multiple stochastic samples of “good” topological network alignments, which allows us to observe regions of the two networks that are robustly aligned across multiple samples. The resulting network alignment frequency (NAF) strongly correlates with GO-based Resnik semantic similarity and enables the first successful cross-species predictions of GO terms based on topology-only network alignments. Our best predictions have an AUPR of about 0.4, which is competitive with state-of-the-art algorithms, even when there is no observable sequence similarity and no known homology relationship. While our results provide only a “proof of concept” on existing network data, we hypothesize that predicting GO terms from topology-only network alignments will become increasingly practical as the volume and quality of PPI network data increase.

Multi-tissue DNA methylation microarray signature is predictive of gene function

ABSTRACT Background Transcriptional correlation networks derived from publicly available gene expression microarrays have been previously shown to be predictive of known gene functions, but less is known about the predictive capacity of correlated DNA methylation at CpG sites. Guilt-by-association co-expression methods can adapted for use with DNA methylation when a representative methylation value is created for each gene. We examine how methylation compares to expression in predicting Gene Ontology terms using both co-methylation and traditional machine learning approaches across different types of representative methylation values per gene. Methods We perform guilt-by-association gene function prediction with a suite of models called Methylation Array Network Analysis, using a network of correlated methylation values derived from over 24,000 samples. In generating the correlation matrix, the performance of different methods of collapsing probe-level data effect on the resulting gene function predictions was compared, along with the use of different regions surrounding the gene of interest. Results Using mean comethylation of a given gene to its annotated term had an overall highest prediction macro-AUC of 0.60 using mean gene body methylation, across all Gene Ontology terms. This was increased using the logistic regression approach with the highest macro-AUC of 0.82 using mean gene body methylation, compared to the naive predictor of 0.72. Conclusion Genes correlated in their methylation state are functionally related. Genes clustered in co-methylation space were enriched for chromatin state, PRC2, immune response, and development-related terms.

HEMDAG: a family of modular and scalable hierarchical ensemble methods to improve Gene Ontology term prediction.

Automated protein function prediction is a complex multi-class, multi-label, structured classification problem in which protein functions are organized in a controlled vocabulary, according to the Gene Ontology (GO). 'Hierarchy-unaware' classifiers, also known as 'flat' methods, predict GO terms without exploiting the inherent structure of the ontology, potentially violating the True-Path-Rule (TPR) that governs the GO, while 'hierarchy-aware' approaches, even if they obey the TPR, do not always show clear improvements with respect to flat methods, or do not scale well when applied to the full GO. To overcome these limitations, we propose Hierarchical Ensemble Methods for Directed Acyclic Graphs (HEMDAG), a family of highly modular hierarchical ensembles of classifiers, able to build upon any flat method and to provide 'TPR-safe' predictions, by leveraging a combination of isotonic regression and TPR learning strategies. Extensive experiments on synthetic and real data across several organisms firstly show that HEMDAG can be used as a general tool to improve the predictions of flat classifiers, and secondly that HEMDAG is competitive versus state-of-the-art hierarchy-aware learning methods proposed in the last CAFA international challenges. Fully tested R code freely available at https://anaconda.org/bioconda/r-hemdag. Tutorial and documentation at https://hemdag.readthedocs.io. Supplementary data are available at Bioinformatics online.

Embeddings from deep learning transfer GO annotations beyond homology

Knowing protein function is crucial to advance molecular and medical biology, yet experimental function annotations through the Gene Ontology (GO) exist for fewer than 0.5% of all known proteins. Computational methods bridge this sequence-annotation gap typically through homology-based annotation transfer by identifying sequence-similar proteins with known function or through prediction methods using evolutionary information. Here, we propose predicting GO terms through annotation transfer based on proximity of proteins in the SeqVec embedding rather than in sequence space. These embeddings originate from deep learned language models (LMs) for protein sequences (SeqVec) transferring the knowledge gained from predicting the next amino acid in 33 million protein sequences. Replicating the conditions of CAFA3, our method reaches an Fmax of 37 ± 2%, 50 ± 3%, and 57 ± 2% for BPO, MFO, and CCO, respectively. Numerically, this appears close to the top ten CAFA3 methods. When restricting the annotation transfer to proteins with < 20% pairwise sequence identity to the query, performance drops (Fmax BPO 33 ± 2%, MFO 43 ± 3%, CCO 53 ± 2%); this still outperforms naïve sequence-based transfer. Preliminary results from CAFA4 appear to confirm these findings. Overall, this new concept is likely to change the annotation of proteins, in particular for proteins from smaller families or proteins with intrinsically disordered regions.

Bioinformatic Prediction of Gene Ontology Terms of Uncharacterized Proteins from Chromosome 11

In chromosome 11, 71 out of its 1254 proteins remain functionally uncharacterized on the basis of their existence evidence (uPE1s) following the latest version of neXtProt (release 2020-01-17). Because in vivo and in vitro experimental strategies are often time-consuming and labor-intensive, there is a need for a bioinformatics tool to predict the function annotation. Here, we used I-TASSER/COFACTOR provided on the neXtProt web site, which predicts gene ontology (GO) terms based on the 3D structure of the protein. I-TASSER/COFACTOR predicted 2413 GO terms with a benchmark dataset of the 22 proteins belonging to PE1 of chromosome 11. In this study, we developed a filtering algorithm in order to select specific GO terms using the GO map generated by I-TASSER/COFACTOR. As a result, 187 specific GO terms showed a higher average precision-recall score at the least cellular component term compared to 2413 predicted GO terms. Next, we applied 65 proteins belonging to uPE1s of chromosome 11, and then 409 out of 6684 GO terms survived, where 103 and 142 GO terms of molecular function and biological process, respectively, were included. Representatively, the cellular component GO terms of CCDC90B, C11orf52, and the SMAP were predicted and validated using the overexpression system into 293T cells and immunofluorescence staining. We will further study their biological and molecular functions toward the goal of the neXt-CP50 project as a part of C-HPP. We shared all results and programs in Github (https://github.com/heeyounh/I-TASSER-COFACTOR-filtering.git).

Blinded Testing of Function Annotation for uPE1 Proteins by I-TASSER/COFACTOR Pipeline Using the 2018-2019 Additions to neXtProt and the CAFA3 Challenge.

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".

Automatic gene annotation using GO terms from cellular component domain

BackgroundThe Gene Ontology (GO) is a resource that supplies information about gene product function using ontologies to represent biological knowledge. These ontologies cover three domains: Cellular Component (CC), Molecular Function (MF), and Biological Process (BP). GO annotation is a process which assigns gene functional information using GO terms to relevant genes in the literature. It is a common task among the Model Organism Database (MOD) groups. Manual GO annotation relies on human curators assigning gene functional information using GO terms by reading the biomedical literature. This process is very time-consuming and labor-intensive. As a result, many MODs can afford to curate only a fraction of relevant articles.MethodsGO terms from the CC domain can be essentially divided into two sub-hierarchies: subcellular location terms, and protein complex terms. We cast the task of gene annotation using GO terms from the CC domain as relation extraction between gene and other entities: (1) extract cases where a protein is found to be in a subcellular location, and (2) extract cases where a protein is a subunit of a protein complex. For each relation extraction task, we use an approach based on triggers and syntactic dependencies to extract the desired relations among entities.ResultsWe tested our approach on the BC4GO test set, a publicly available corpus for GO annotation. Our approach obtains a F1-score of 71%, a precision of 91% and a recall of 58% for predicting GO terms from CC Domain for given genes.ConclusionsWe have described a novel approach of treating gene annotation with GO terms from CC domain as two relation extraction subtasks. Evaluation results show that our approach achieves a F1-score of 71% for predicting GO terms for given genes. Thereby our approach can be used to accelerate the process of GO annotation for the bio-annotators.

Evaluating the impact of topological protein features on the negative examples selection

BackgroundSupervised machine learning methods when applied to the problem of automated protein-function prediction (AFP) require the availability of both positive examples (i.e., proteins which are known to possess a given protein function) and negative examples (corresponding to proteins not associated with that function). Unfortunately, publicly available proteome and genome data sources such as the Gene Ontology rarely store the functions not possessed by a protein. Thus the negative selection, consisting in identifying informative negative examples, is currently a central and challenging problem in AFP. Several heuristics have been proposed through the years to solve this problem; nevertheless, despite their effectiveness, to the best of our knowledge no previous existing work studied which protein features are more relevant to this task, that is, which protein features help more in discriminating reliable and unreliable negatives.ResultsThe present work analyses the impact of several features on the selection of negative proteins for the Gene Ontology (GO) terms. The analysis is network-based: it exploits the fact that proteins can be naturally structured in a network, considering the pairwise relationships coming from several sources of data, such as protein-protein and genetic interactions. Overall, the proposed protein features, including local and global graph centrality measures and protein multifunctionality, can be term-aware (i.e., depending on the GO term) and term-unaware (i.e., invariant across the GO terms). We validated the informativeness of each feature utilizing a temporal holdout in three different experiments on yeast, mouse and human proteomes: (i) feature selection to detect which protein features are more helpful for the negative selection; (ii) protein function prediction to verify whether the features considered are also useful to predict GO terms; (iii) negative selection by applying two different negative selection algorithms on proteins represented through the proposed features.ConclusionsTerm-aware features (with some exceptions) resulted more informative for problem (i), together with node betweenness, which is the most relevant among term-unaware features. The node positive neighborhood instead is the most predictive feature for the AFP problem, while experiment (iii) showed that the proposed features allow negative selection algorithms to select effectively negative instances in the temporal holdout setting, with better results when nonlinear combinations of features are also exploited.

Applicability of semi-supervised learning assumptions for gene ontology terms prediction

Gene Ontology (GO) is one of the most important resources in bioinformatics, aiming to provide a unified framework for the biological annotation of genes and proteins across all species. Predicting GO terms is an essential task for bioinformatics, but the number of available labelled proteins is in several cases insufficient for training reliable machine learning classifiers. Semi-supervised learning methods arise as a powerful solution that explodes the information contained in unlabelled data in order to improve the estimations of traditional supervised approaches. However, semi-supervised learning methods have to make strong assumptions about the nature of the training data and thus, the performance of the predictor is highly dependent on these assumptions. This paper presents an analysis of the applicability of semi-supervised learning assumptions over the specific task of GO terms prediction, focused on providing judgment elements that allow choosing the most suitable tools for specific GO terms. The results show that semi-supervised approaches significantly outperform the traditional supervised methods and that the highest performances are reached when applying the cluster assumption. Besides, it is experimentally demonstrated that cluster and manifold assumptions are complementary to each other and an analysis of which GO terms can be more prone to be correctly predicted with each assumption, is provided.

Predicting gene ontology annotations of orphan GWAS genes using protein-protein interactions

BackgroundThe number of genome-wide association studies (GWAS) has increased rapidly in the past couple of years, resulting in the identification of genes associated with different diseases. The next step in translating these findings into biomedically useful information is to find out the mechanism of the action of these genes. However, GWAS studies often implicate genes whose functions are currently unknown; for example, MYEOV, ANKLE1, TMEM45B and ORAOV1 are found to be associated with breast cancer, but their molecular function is unknown.ResultsWe carried out Bayesian inference of Gene Ontology (GO) term annotations of genes by employing the directed acyclic graph structure of GO and the network of protein-protein interactions (PPIs). The approach is designed based on the fact that two proteins that interact biophysically would be in physical proximity of each other, would possess complementary molecular function, and play role in related biological processes. Predicted GO terms were ranked according to their relative association scores and the approach was evaluated quantitatively by plotting the precision versus recall values and F-scores (the harmonic mean of precision and recall) versus varying thresholds. Precisions of ~58% and ~ 40% for localization and functions respectively of proteins were determined at a threshold of ~30 (top 30 GO terms in the ranked list). Comparison with function prediction based on semantic similarity among nodes in an ontology and incorporation of those similarities in a k-nearest neighbor classifier confirmed that our results compared favorably.ConclusionsThis approach was applied to predict the cellular component and molecular function GO terms of all human proteins that have interacting partners possessing at least one known GO annotation. The list of predictions is available at http://severus.dbmi.pitt.edu/engo/GOPRED.html. We present the algorithm, evaluations and the results of the computational predictions, especially for genes identified in GWAS studies to be associated with diseases, which are of translational interest.

Predictability of gene ontology slim-terms from primary structure information in Embryophyta plant proteins

BackgroundProteins are the key elements on the path from genetic information to the development of life. The roles played by the different proteins are difficult to uncover experimentally as this process involves complex procedures such as genetic modifications, injection of fluorescent proteins, gene knock-out methods and others. The knowledge learned from each protein is usually annotated in databases through different methods such as the proposed by The Gene Ontology (GO) consortium. Different methods have been proposed in order to predict GO terms from primary structure information, but very few are available for large-scale functional annotation of plants, and reported success rates are much less than the reported by other non-plant predictors. This paper explores the predictability of GO annotations on proteins belonging to the Embryophyta group from a set of features extracted solely from their primary amino acid sequence.ResultsHigh predictability of several GO terms was found for Molecular Function and Cellular Component. As expected, a lower degree of predictability was found on Biological Process ontology annotations, although a few biological processes were easily predicted. Proteins related to transport and transcription were particularly well predicted from primary structure information. The most discriminant features for prediction were those related to electric charges of the amino-acid sequence and hydropathicity derived features.ConclusionsAn analysis of GO-slim terms predictability in plants was carried out, in order to determine single categories or groups of functions that are most related with primary structure information. For each highly predictable GO term, the responsible features of such successfulness were identified and discussed. In addition to most published studies, focused on few categories or single ontologies, results in this paper comprise a complete landscape of GO predictability from primary structure encompassing 75 GO terms at molecular, cellular and phenotypical level. Thus, it provides a valuable guide for researchers interested on further advances in protein function prediction on Embryophyta plants.

Assigning roles to DNA regulatory motifs using comparative genomics

Motivation: Transcription factors (TFs) are crucial during the lifetime of the cell. Their functional roles are defined by the genes they regulate. Uncovering these roles not only sheds light on the TF at hand but puts it into the context of the complete regulatory network.Results: Here, we present an alignment- and threshold-free comparative genomics approach for assigning functional roles to DNA regulatory motifs. We incorporate our approach into the Gomo algorithm, a computational tool for detecting associations between a user-specified DNA regulatory motif [expressed as a position weight matrix (PWM)] and Gene Ontology (GO) terms. Incorporating multiple species into the analysis significantly improves Gomo's ability to identify GO terms associated with the regulatory targets of TFs. Including three comparative species in the process of predicting TF roles in Saccharomyces cerevisiae and Homo sapiens increases the number of significant predictions by 75 and 200%, respectively. The predicted GO terms are also more specific, yielding deeper biological insight into the role of the TF. Adjusting motif (binding) affinity scores for individual sequence composition proves to be essential for avoiding false positive associations. We describe a novel DNA sequence-scoring algorithm that compensates a thermodynamic measure of DNA-binding affinity for individual sequence base composition. Gomo's prediction accuracy proves to be relatively insensitive to how promoters are defined. Because Gomo uses a threshold-free form of gene set analysis, there are no free parameters to tune. Biologists can investigate the potential roles of DNA regulatory motifs of interest using Gomo via the web (http://meme.nbcr.net).Contact: t.bailey@uq.edu.auSupplementary information: Supplementary data are available at Bioinformatics online.

Associating transcription factor-binding site motifs with target GO terms and target genes

The roles and target genes of many transcription factors (TFs) are still unknown. To predict the roles of TFs, we present a computational method for associating Gene Ontology (GO) terms with TF-binding motifs. The method works by ranking all genes as potential targets of the TF, and reporting GO terms that are significantly associated with highly ranked genes. We also present an approach, whereby these predicted GO terms can be used to improve predictions of TF target genes. This uses a novel gene-scoring function that reflects the insight that genes annotated with GO terms predicted to be associated with the TF are more likely to be its targets. We construct validation sets of GO terms highly associated with known targets of various yeast and human TF. On the yeast reference sets, our prediction method identifies at least one correct GO term for 73% of the TF, 49% of the correct GO terms are predicted and almost one-third of the predicted GO terms are correct. Results on human reference sets are similarly encouraging. Validation of our target gene prediction method shows that its accuracy exceeds that of simple motif scanning.

Consistent probabilistic outputs for protein function prediction

In predicting hierarchical protein function annotations, such as terms in the Gene Ontology (GO), the simplest approach makes predictions for each term independently. However, this approach has the unfortunate consequence that the predictor may assign to a single protein a set of terms that are inconsistent with one another; for example, the predictor may assign a specific GO term to a given protein ('purine nucleotide binding') but not assign the parent term ('nucleotide binding'). Such predictions are difficult to interpret. In this work, we focus on methods for calibrating and combining independent predictions to obtain a set of probabilistic predictions that are consistent with the topology of the ontology. We call this procedure 'reconciliation'. We begin with a baseline method for predicting GO terms from a collection of data types using an ensemble of discriminative classifiers. We apply the method to a previously described benchmark data set, and we demonstrate that the resulting predictions are frequently inconsistent with the topology of the GO. We then consider 11 distinct reconciliation methods: three heuristic methods; four variants of a Bayesian network; an extension of logistic regression to the structured case; and three novel projection methods - isotonic regression and two variants of a Kullback-Leibler projection method. We evaluate each method in three different modes - per term, per protein and joint - corresponding to three types of prediction tasks. Although the principal goal of reconciliation is interpretability, it is important to assess whether interpretability comes at a cost in terms of precision and recall. Indeed, we find that many apparently reasonable reconciliation methods yield reconciled probabilities with significantly lower precision than the original, unreconciled estimates. On the other hand, we find that isotonic regression usually performs better than the underlying, unreconciled method, and almost never performs worse; isotonic regression appears to be able to use the constraints from the GO network to its advantage. An exception to this rule is the high precision regime for joint evaluation, where Kullback-Leibler projection yields the best performance.

An en masse phenotype and function prediction system for Mus musculus

Background:Individual researchers are struggling to keep up with the accelerating emergence of high-throughput biological data, and to extract information that relates to their specific questions. Integration of accumulated evidence should permit researchers to form fewer - and more accurate - hypotheses for further study through experimentation.Results:Here a method previously used to predict Gene Ontology (GO) terms for Saccharomyces cerevisiae (Tian et al.: Combining guilt-by-association and guilt-by-profiling to predict Saccharomyces cerevisiae gene function. Genome Biol 2008, 9(Suppl 1):S7) is applied to predict GO terms and phenotypes for 21,603 Mus musculus genes, using a diverse collection of integrated data sources (including expression, interaction, and sequence-based data). This combined 'guilt-by-profiling' and 'guilt-by-association' approach optimizes the combination of two inference methodologies. Predictions at all levels of confidence are evaluated by examining genes not used in training, and top predictions are examined manually using available literature and knowledge base resources.Conclusion:We assigned a confidence score to each gene/term combination. The results provided high prediction performance, with nearly every GO term achieving greater than 40% precision at 1% recall. Among the 36 novel predictions for GO terms and 40 for phenotypes that were studied manually, >80% and >40%, respectively, were identified as accurate. We also illustrate that a combination of 'guilt-by-profiling' and 'guilt-by-association' outperforms either approach alone in their application to M. musculus.