NetGO: improving large-scale protein function prediction with massive network information

Deep learning methods for protein function prediction

Predicting protein function from protein sequence, structure, interaction, and other relevant information is important for generating hypotheses for biological experiments and studying biological systems, and therefore has been a major challenge in protein bioinformatics. Numerous computational methods had been developed to advance protein function prediction gradually in the last two decades. Particularly, in the recent years, leveraging the revolutionary advances in artificial intelligence (AI), more and more deep learning methods have been developed to improve protein function prediction at a faster pace. Here, we provide an in-depth review of the recent developments of deep learning methods for protein function prediction. We summarize the significant advances in the field, identify several remaining major challenges to be tackled, and suggest some potential directions to explore. The data sources and evaluation metrics widely used in protein function prediction are also discussed to assist the machine learning, AI, and bioinformatics communities to develop more cutting-edge methods to advance protein function prediction.

GPSFun: geometry-aware protein sequence function predictions with language models

Knowledge of protein function is essential for elucidating disease mechanisms and discovering new drug targets. However, there is a widening gap between the exponential growth of protein sequences and their limited function annotations. In our prior studies, we have developed a series of methods including GraphPPIS, GraphSite, LMetalSite and SPROF-GO for protein function annotations at residue or protein level. To further enhance their applicability and performance, we now present GPSFun, a versatile web server for Geometry-aware Protein Sequence Function annotations, which equips our previous tools with language models and geometric deep learning. Specifically, GPSFun employs large language models to efficiently predict 3D conformations of the input protein sequences and extract informative sequence embeddings. Subsequently, geometric graph neural networks are utilized to capture the sequence and structure patterns in the protein graphs, facilitating various downstream predictions including protein-ligand binding sites, gene ontologies, subcellular locations and protein solubility. Notably, GPSFun achieves superior performance to state-of-the-art methods across diverse tasks without requiring multiple sequence alignments or experimental protein structures. GPSFun is freely available to all users at https://bio-web1.nscc-gz.cn/app/GPSFun with user-friendly interfaces and rich visualizations.

POLAT: Protein function prediction based on soft mask graph network and residue-Label ATtention

MOTIVATION

Elucidating protein function is a central problem in biochemistry, genetics, and molecular biology. Developing computational methods for protein function prediction is critical due to the significant gap between sequence and functional data. Recent advances in protein structure prediction, which strongly correlates with function, make it feasible to use structure to predict function. However, current structure-based methods overlook the fact that individual residues may contribute differently to the protein's function and do not take into account the correlation between protein residues and their functions. The challenge of effectively utilizing the relationship between protein residues and function-level information to predict protein function remains unsolved.

RESULT

We proposed a protein function prediction method based on Soft Mask Graph Networks and Residue-Label Attention (POLAT), which could combine sequence features, predicted structure features, and function-level information to get an accurate prediction. We use soft mask graph networks to adaptively extract the residues relevant to functions. A residue-label attention mechanism is adopted to obtain the protein-level encoded features of a protein, which are then concatenated with a protein-level embedding and fed into a dense classifier to determine the probabilities of each function. POLAT achieves 0.670, 0.515, 0.578 Fmax and 0.677, 0.409, 0.507 AUPR on the PDB cdhit test set for the MFO, BPO, and CCO domains, respectively, outperforming the existing structure-based SOTA method GAT-GO (Fmax 0.633, 0.492, 0.547; AUPR 0.660, 0.381, 0.479). POLAT is also competitive in extensive experiments among sequence-based and multimodal methods and achieves the SOTA performance in three out of six metrics.

Protein function prediction through multi-view multi-label latent tensor reconstruction

BACKGROUND

In last two decades, the use of high-throughput sequencing technologies has accelerated the pace of discovery of proteins. However, due to the time and resource limitations of rigorous experimental functional characterization, the functions of a vast majority of them remain unknown. As a result, computational methods offering accurate, fast and large-scale assignment of functions to new and previously unannotated proteins are sought after. Leveraging the underlying associations between the multiplicity of features that describe proteins could reveal functional insights into the diverse roles of proteins and improve performance on the automatic function prediction task.

RESULTS

We present GO-LTR, a multi-view multi-label prediction model that relies on a high-order tensor approximation of model weights combined with non-linear activation functions. The model is capable of learning high-order relationships between multiple input views representing the proteins and predicting high-dimensional multi-label output consisting of protein functional categories. We demonstrate the competitiveness of our method on various performance measures. Experiments show that GO-LTR learns polynomial combinations between different protein features, resulting in improved performance. Additional investigations establish GO-LTR's practical potential in assigning functions to proteins under diverse challenging scenarios: very low sequence similarity to previously observed sequences, rarely observed and highly specific terms in the gene ontology.

IMPLEMENTATION

The code and data used for training GO-LTR is available at https://github.com/aalto-ics-kepaco/GO-LTR-prediction .

Exploring salt tolerance mechanisms using machine learning for transcriptomic insights: case study in Spartina alterniflora

Salt stress poses a significant threat to global cereal crop production, emphasizing the need for a comprehensive understanding of salt tolerance mechanisms. Accurate functional annotations of differentially expressed genes are crucial for gaining insights into the salt tolerance mechanism. The challenge of predicting gene functions in under-studied species, especially when excluding infrequent GO terms, persists. Therefore, we proposed the use of NetGO 3.0, a machine learning-based annotation method that does not rely on homology information between species, to predict the functions of differentially expressed genes under salt stress. Spartina alterniflora, a halophyte with salt glands, exhibits remarkable salt tolerance, making it an excellent candidate for in-depth transcriptomic analysis. However, current research on the S. alterniflora transcriptome under salt stress is limited. In this study we used S. alterniflora as an example to investigate its transcriptional responses to various salt concentrations, with a focus on understanding its salt tolerance mechanisms. Transcriptomic analysis revealed substantial changes impacting key pathways, such as gene transcription, ion transport, and ROS metabolism. Notably, we identified a member of the SWEET gene family in S. alterniflora, SA_12G129900.m1, showing convergent selection with the rice ortholog SWEET15. Additionally, our genome-wide analyses explored alternative splicing responses to salt stress, providing insights into the parallel functions of alternative splicing and transcriptional regulation in enhancing salt tolerance in S. alterniflora. Surprisingly, there was minimal overlap between differentially expressed and differentially spliced genes following salt exposure. This innovative approach, combining transcriptomic analysis with machine learning-based annotation, avoids the reliance on homology information and facilitates the discovery of unknown gene functions, and is applicable across all sequenced species.

A Comprehensive Review on Machine Learning Techniques for Protein Family Prediction

Proteomics is a field dedicated to the analysis of proteins in cells, tissues, and organisms, aiming to gain insights into their structures, functions, and interactions. A crucial aspect within proteomics is protein family prediction, which involves identifying evolutionary relationships between proteins by examining similarities in their sequences or structures. This approach holds great potential for applications such as drug discovery and functional annotation of genomes. However, current methods for protein family prediction have certain limitations, including limited accuracy, high false positive rates, and challenges in handling large datasets. Some methods also rely on homologous sequences or protein structures, which introduce biases and restrict their applicability to specific protein families or structures. To overcome these limitations, researchers have turned to machine learning (ML) approaches that can identify connections between protein features and simplify complex high-dimensional datasets. This paper presents a comprehensive survey of articles that employ various ML techniques for predicting protein families. The primary objective is to explore and improve ML techniques specifically for protein family prediction, thus advancing future research in the field. Through qualitative and quantitative analyses of ML techniques, it is evident that multiple methods utilizing a range of classifiers have been applied for protein family prediction. However, there has been limited focus on developing novel classifiers for protein family classification, highlighting the urgent need for improved approaches in this area. By addressing these challenges, this research aims to enhance the accuracy and effectiveness of protein family prediction, ultimately facilitating advancements in proteomics and its diverse applications.

PLMSearch: Protein language model powers accurate and fast sequence search for remote homology

Homologous protein search is one of the most commonly used methods for protein annotation and analysis. Compared to structure search, detecting distant evolutionary relationships from sequences alone remains challenging. Here we propose PLMSearch (Protein Language Model), a homologous protein search method with only sequences as input. PLMSearch uses deep representations from a pre-trained protein language model and trains the similarity prediction model with a large number of real structure similarity. This enables PLMSearch to capture the remote homology information concealed behind the sequences. Extensive experimental results show that PLMSearch can search millions of query-target protein pairs in seconds like MMseqs2 while increasing the sensitivity by more than threefold, and is comparable to state-of-the-art structure search methods. In particular, unlike traditional sequence search methods, PLMSearch can recall most remote homology pairs with dissimilar sequences but similar structures. PLMSearch is freely available at https://dmiip.sjtu.edu.cn/PLMSearch .

DeepSS2GO: protein function prediction from secondary structure

Predicting protein function is crucial for understanding biological life processes, preventing diseases and developing new drug targets. In recent years, methods based on sequence, structure and biological networks for protein function annotation have been extensively researched. Although obtaining a protein in three-dimensional structure through experimental or computational methods enhances the accuracy of function prediction, the sheer volume of proteins sequenced by high-throughput technologies presents a significant challenge. To address this issue, we introduce a deep neural network model DeepSS2GO (Secondary Structure to Gene Ontology). It is a predictor incorporating secondary structure features along with primary sequence and homology information. The algorithm expertly combines the speed of sequence-based information with the accuracy of structure-based features while streamlining the redundant data in primary sequences and bypassing the time-consuming challenges of tertiary structure analysis. The results show that the prediction performance surpasses state-of-the-art algorithms. It has the ability to predict key functions by effectively utilizing secondary structure information, rather than broadly predicting general Gene Ontology terms. Additionally, DeepSS2GO predicts five times faster than advanced algorithms, making it highly applicable to massive sequencing data. The source code and trained models are available at https://github.com/orca233/DeepSS2GO.

Multi-Omics integration can be used to rescue metabolic information for some of the dark region of the Pseudomonas putida proteome

In every omics experiment, genes or their products are identified for which even state of the art tools are unable to assign a function. In the biotechnology chassis organism Pseudomonas putida, these proteins of unknown function make up 14% of the proteome. This missing information can bias analyses since these proteins can carry out functions which impact the engineering of organisms. As a consequence of predicting protein function across all organisms, function prediction tools generally fail to use all of the types of data available for any specific organism, including protein and transcript expression information. Additionally, the release of Alphafold predictions for all Uniprot proteins provides a novel opportunity for leveraging structural information. We constructed a bespoke machine learning model to predict the function of recalcitrant proteins of unknown function in Pseudomonas putida based on these sources of data, which annotated 1079 terms to 213 proteins. Among the predicted functions supplied by the model, we found evidence for a significant overrepresentation of nitrogen metabolism and macromolecule processing proteins. These findings were corroborated by manual analyses of selected proteins which identified, among others, a functionally unannotated operon that likely encodes a branch of the shikimate pathway.

AnnoPRO: a strategy for protein function annotation based on multi-scale protein representation and a hybrid deep learning of dual-path encoding

Protein function annotation has been one of the longstanding issues in biological sciences, and various computational methods have been developed. However, the existing methods suffer from a serious long-tail problem, with a large number of GO families containing few annotated proteins. Herein, an innovative strategy named AnnoPRO was therefore constructed by enabling sequence-based multi-scale protein representation, dual-path protein encoding using pre-training, and function annotation by long short-term memory-based decoding. A variety of case studies based on different benchmarks were conducted, which confirmed the superior performance of AnnoPRO among available methods. Source code and models have been made freely available at: https://github.com/idrblab/AnnoPRO and https://zenodo.org/records/10012272.

Numerical stability of DeepGOPlus inference

Convolutional neural networks (CNNs) are currently among the most widely-used deep neural network (DNN) architectures available and achieve state-of-the-art performance for many problems. Originally applied to computer vision tasks, CNNs work well with any data with a spatial relationship, besides images, and have been applied to different fields. However, recent works have highlighted numerical stability challenges in DNNs, which also relates to their known sensitivity to noise injection. These challenges can jeopardise their performance and reliability. This paper investigates DeepGOPlus, a CNN that predicts protein function. DeepGOPlus has achieved state-of-the-art performance and can successfully take advantage and annotate the abounding protein sequences emerging in proteomics. We determine the numerical stability of the model's inference stage by quantifying the numerical uncertainty resulting from perturbations of the underlying floating-point data. In addition, we explore the opportunity to use reduced-precision floating point formats for DeepGOPlus inference, to reduce memory consumption and latency. This is achieved by instrumenting DeepGOPlus' execution using Monte Carlo Arithmetic, a technique that experimentally quantifies floating point operation errors and VPREC, a tool that emulates results with customizable floating point precision formats. Focus is placed on the inference stage as it is the primary deliverable of the DeepGOPlus model, widely applicable across different environments. All in all, our results show that although the DeepGOPlus CNN is very stable numerically, it can only be selectively implemented with lower-precision floating-point formats. We conclude that predictions obtained from the pre-trained DeepGOPlus model are very reliable numerically, and use existing floating-point formats efficiently.

A comprehensive computational benchmark for evaluating deep learning-based protein function prediction approaches

Proteins play an important role in life activities and are the basic units for performing functions. Accurately annotating functions to proteins is crucial for understanding the intricate mechanisms of life and developing effective treatments for complex diseases. Traditional biological experiments struggle to keep pace with the growing number of known proteins. With the development of high-throughput sequencing technology, a wide variety of biological data provides the possibility to accurately predict protein functions by computational methods. Consequently, many computational methods have been proposed. Due to the diversity of application scenarios, it is necessary to conduct a comprehensive evaluation of these computational methods to determine the suitability of each algorithm for specific cases. In this study, we present a comprehensive benchmark, BeProf, to process data and evaluate representative computational methods. We first collect the latest datasets and analyze the data characteristics. Then, we investigate and summarize 17 state-of-the-art computational methods. Finally, we propose a novel comprehensive evaluation metric, design eight application scenarios and evaluate the performance of existing methods on these scenarios. Based on the evaluation, we provide practical recommendations for different scenarios, enabling users to select the most suitable method for their specific needs. All of these servers can be obtained from https://csuligroup.com/BEPROF and https://github.com/CSUBioGroup/BEPROF.

A CNN-CBAM-BIGRU model for protein function prediction

Understanding a protein's function based solely on its amino acid sequence is a crucial but intricate task in bioinformatics. Traditionally, this challenge has proven difficult. However, recent years have witnessed the rise of deep learning as a powerful tool, achieving significant success in protein function prediction. Their strength lies in their ability to automatically learn informative features from protein sequences, which can then be used to predict the protein's function. This study builds upon these advancements by proposing a novel model: CNN-CBAM+BiGRU. It incorporates a Convolutional Block Attention Module (CBAM) alongside BiGRUs. CBAM acts as a spotlight, guiding the CNN to focus on the most informative parts of the protein data, leading to more accurate feature extraction. BiGRUs, a type of Recurrent Neural Network (RNN), excel at capturing long-range dependencies within the protein sequence, which are essential for accurate function prediction. The proposed model integrates the strengths of both CNN-CBAM and BiGRU. This study's findings, validated through experimentation, showcase the effectiveness of this combined approach. For the human dataset, the suggested method outperforms the CNN-BIGRU+ATT model by +1.0 % for cellular components, +1.1 % for molecular functions, and +0.5 % for biological processes. For the yeast dataset, the suggested method outperforms the CNN-BIGRU+ATT model by +2.4 % for the cellular component, +1.2 % for molecular functions, and +0.6 % for biological processes.

FURNA: a database for function annotations of RNA structures

Despite the increasing number of 3D RNA structures in the Protein Data Bank, the majority of experimental RNA structures lack thorough functional annotations. As the significance of the functional roles played by non-coding RNAs becomes increasingly apparent, comprehensive annotation of RNA function is becoming a pressing concern. In response to this need, we have developed FURNA (Functions of RNAs), the first database for experimental RNA structures that aims to provide a comprehensive repository of high-quality functional annotations. These include Gene Ontology terms, Enzyme Commission numbers, ligand binding sites, RNA families, protein binding motifs, and cross-references to related databases. FURNA is available at https://seq2fun.dcmb.med.umich.edu/furna/ to enable quick discovery of RNA functions from their structures and sequences.

Ensemble Learning with Supervised Methods Based on Large-Scale Protein Language Models for Protein Mutation Effects Prediction

Machine learning has been increasingly utilized in the field of protein engineering, and research directed at predicting the effects of protein mutations has attracted increasing attention. Among them, so far, the best results have been achieved by related methods based on protein language models, which are trained on a large number of unlabeled protein sequences to capture the generally hidden evolutionary rules in protein sequences, and are therefore able to predict their fitness from protein sequences. Although numerous similar models and methods have been successfully employed in practical protein engineering processes, the majority of the studies have been limited to how to construct more complex language models to capture richer protein sequence feature information and utilize this feature information for unsupervised protein fitness prediction. There remains considerable untapped potential in these developed models, such as whether the prediction performance can be further improved by integrating different models to further improve the accuracy of prediction. Furthermore, how to utilize large-scale models for prediction methods of mutational effects on quantifiable properties of proteins due to the nonlinear relationship between protein fitness and the quantification of specific functionalities has yet to be explored thoroughly. In this study, we propose an ensemble learning approach for predicting mutational effects of proteins integrating protein sequence features extracted from multiple large protein language models, as well as evolutionarily coupled features extracted in homologous sequences, while comparing the differences between linear regression and deep learning models in mapping these features to quantifiable functional changes. We tested our approach on a dataset of 17 protein deep mutation scans and indicated that the integrated approach together with linear regression enables the models to have higher prediction accuracy and generalization. Moreover, we further illustrated the reliability of the integrated approach by exploring the differences in the predictive performance of the models across species and protein sequence lengths, as well as by visualizing clustering of ensemble and non-ensemble features.

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. 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. This study emphasizes the critical role of search parameter settings in homology-based function transfer.

HNetGO: protein function prediction via heterogeneous network transformer

Protein function annotation is one of the most important research topics for revealing the essence of life at molecular level in the post-genome era. Current research shows that integrating multisource data can effectively improve the performance of protein function prediction models. However, the heavy reliance on complex feature engineering and model integration methods limits the development of existing methods. Besides, models based on deep learning only use labeled data in a certain dataset to extract sequence features, thus ignoring a large amount of existing unlabeled sequence data. Here, we propose an end-to-end protein function annotation model named HNetGO, which innovatively uses heterogeneous network to integrate protein sequence similarity and protein-protein interaction network information and combines the pretraining model to extract the semantic features of the protein sequence. In addition, we design an attention-based graph neural network model, which can effectively extract node-level features from heterogeneous networks and predict protein function by measuring the similarity between protein nodes and gene ontology term nodes. Comparative experiments on the human dataset show that HNetGO achieves state-of-the-art performance on cellular component and molecular function branches.

Prot2GO: Predicting GO Annotations From Protein Sequences and Interactions

Protein is the main material basis of living organisms and plays crucial role in life activities. Understanding the function of protein is of great significance for new drug discovery, disease treatment and vaccine development. In recent years, with the widespread application of deep learning in bioinformatics, researchers have proposed many deep learning models to predict protein functions. However, the existing deep learning methods usually only consider protein sequences, and thus cannot effectively integrate multi-source data to annotate protein functions. In this article, we propose the Prot2GO model, which can integrate protein sequence and PPI network data to predict protein functions. We utilize an improved biased random walk algorithm to extract the features of PPI network. For sequence data, we use a convolutional neural network to obtain the local features of the sequence and a recurrent neural network to capture the long-range associations between amino acid residues in protein sequence. Moreover, Prot2GO adopts the attention mechanism to identify protein motifs and structural domains. Experiments show that Prot2GO model achieves the state-of-the-art performance on multiple metrics.

Large-scale predicting protein functions through heterogeneous feature fusion

As the volume of protein sequence and structure data grows rapidly, the functions of the overwhelming majority of proteins cannot be experimentally determined. Automated annotation of protein function at a large scale is becoming increasingly important. Existing computational prediction methods are typically based on expanding the relatively small number of experimentally determined functions to large collections of proteins with various clues, including sequence homology, protein-protein interaction, gene co-expression, etc. Although there has been some progress in protein function prediction in recent years, the development of accurate and reliable solutions still has a long way to go. Here we exploit AlphaFold predicted three-dimensional structural information, together with other non-structural clues, to develop a large-scale approach termed PredGO to annotate Gene Ontology (GO) functions for proteins. We use a pre-trained language model, geometric vector perceptrons and attention mechanisms to extract heterogeneous features of proteins and fuse these features for function prediction. The computational results demonstrate that the proposed method outperforms other state-of-the-art approaches for predicting GO functions of proteins in terms of both coverage and accuracy. The improvement of coverage is because the number of structures predicted by AlphaFold is greatly increased, and on the other hand, PredGO can extensively use non-structural information for functional prediction. Moreover, we show that over 205 000 ($\sim $100%) entries in UniProt for human are annotated by PredGO, over 186 000 ($\sim $90%) of which are based on predicted structure. The webserver and database are available at http://predgo.denglab.org/.

Combining protein sequences and structures with transformers and equivariant graph neural networks to predict protein function

MOTIVATION

Millions of protein sequences have been generated by numerous genome and transcriptome sequencing projects. However, experimentally determining the function of the proteins is still a time consuming, low-throughput, and expensive process, leading to a large protein sequence-function gap. Therefore, it is important to develop computational methods to accurately predict protein function to fill the gap. Even though many methods have been developed to use protein sequences as input to predict function, much fewer methods leverage protein structures in protein function prediction because there was lack of accurate protein structures for most proteins until recently.

RESULTS

We developed TransFun-a method using a transformer-based protein language model and 3D-equivariant graph neural networks to distill information from both protein sequences and structures to predict protein function. It extracts feature embeddings from protein sequences using a pre-trained protein language model (ESM) via transfer learning and combines them with 3D structures of proteins predicted by AlphaFold2 through equivariant graph neural networks. Benchmarked on the CAFA3 test dataset and a new test dataset, TransFun outperforms several state-of-the-art methods, indicating that the language model and 3D-equivariant graph neural networks are effective methods to leverage protein sequences and structures to improve protein function prediction. Combining TransFun predictions and sequence similarity-based predictions can further increase prediction accuracy.

AVAILABILITY AND IMPLEMENTATION

The source code of TransFun is available at https://github.com/jianlin-cheng/TransFun.

NetGO 3.0: Protein Language Model Improves Large-scale Functional Annotations

As one of the state-of-the-art automated function prediction (AFP) methods, NetGO 2.0 integrates multi-source information to improve the performance. However, it mainly utilizes the proteins with experimentally supported functional annotations without leveraging valuable information from a vast number of unannotated proteins. Recently, protein language models have been proposed to learn informative representations [e.g., Evolutionary Scale Modeling (ESM)-1b embedding] from protein sequences based on self-supervision. Here, we represented each protein by ESM-1b and used logistic regression (LR) to train a new model, LR-ESM, for AFP. The experimental results showed that LR-ESM achieved comparable performance with the best-performing component of NetGO 2.0. Therefore, by incorporating LR-ESM into NetGO 2.0, we developed NetGO 3.0 to improve the performance of AFP extensively. NetGO 3.0 is freely accessible at https://dmiip.sjtu.edu.cn/ng3.0.

ProFAB-open protein functional annotation benchmark

As the number of protein sequences increases in biological databases, computational methods are required to provide accurate functional annotation with high coverage. Although several machine learning methods have been proposed for this purpose, there are still two main issues: (i) construction of reliable positive and negative training and validation datasets, and (ii) fair evaluation of their performances based on predefined experimental settings. To address these issues, we have developed ProFAB: Open Protein Functional Annotation Benchmark, which is a platform providing an infrastructure for a fair comparison of protein function prediction methods. ProFAB provides filtered and preprocessed protein annotation datasets and enables the training and evaluation of function prediction methods via several options. We believe that ProFAB will be useful for both computational and experimental researchers by enabling the utilization of ready-to-use datasets and machine learning algorithms for protein function prediction based on Gene Ontology terms and Enzyme Commission numbers. ProFAB is available at https://github.com/kansil/ProFAB and https://profab.kansil.org.

A Deep Learning Framework for Predicting Protein Functions With Co-Occurrence of GO Terms

The understanding of protein functions is critical to many biological problems such as the development of new drugs and new crops. To reduce the huge gap between the increase of protein sequences and annotations of protein functions, many methods have been proposed to deal with this problem. These methods use Gene Ontology (GO) to classify the functions of proteins and consider one GO term as a class label. However, they ignore the co-occurrence of GO terms that is helpful for protein function prediction. We propose a new deep learning model, named DeepPFP-CO, which uses Graph Convolutional Network (GCN) to explore and capture the co-occurrence of GO terms to improve the protein function prediction performance. In this way, we can further deduce the protein functions by fusing the predicted propensity of the center function and its co-occurrence functions. We use Fmax and AUPR to evaluate the performance of DeepPFP-CO and compare DeepPFP-CO with state-of-the-art methods such as DeepGOPlus and DeepGOA. The computational results show that DeepPFP-CO outperforms DeepGOPlus and other methods. Moreover, we further analyze our model at the protein level. The results have demonstrated that DeepPFP-CO improves the performance of protein function prediction. DeepPFP-CO is available at https://csuligroup.com/DeepPFP/.

A systematic review of state-of-the-art strategies for machine learning-based protein function prediction

New drug discovery is inseparable from the discovery of drug targets, and the vast majority of the known targets are proteins. At the same time, proteins are essential structural and functional elements of living cells necessary for the maintenance of all forms of life. Therefore, protein functions have become the focus of many pharmacological and biological studies. Traditional experimental techniques are no longer adequate for rapidly growing annotation of protein sequences, and approaches to protein function prediction using computational methods have emerged and flourished. A significant trend has been to use machine learning to achieve this goal. In this review, approaches to protein function prediction based on the sequence, structure, protein-protein interaction (PPI) networks, and fusion of multi-information sources are discussed. The current status of research on protein function prediction using machine learning is considered, and existing challenges and prominent breakthroughs are discussed to provide ideas and methods for future studies.

Combining protein sequences and structures with transformers and equivariant graph neural networks to predict protein function

Motivation

Millions of protein sequences have been generated by numerous genome and transcriptome sequencing projects. However, experimentally determining the function of the proteins is still a time consuming, low-throughput, and expensive process, leading to a large protein sequence-function gap. Therefore, it is important to develop computational methods to accurately predict protein function to fill the gap. Even though many methods have been developed to use protein sequences as input to predict function, much fewer methods leverage protein structures in protein function prediction because there was lack of accurate protein structures for most proteins until recently.

Results

We developed TransFun - a method using a transformer-based protein language model and 3D-equivariant graph neural networks to distill information from both protein sequences and structures to predict protein function. It extracts feature embeddings from protein sequences using a pre-trained protein language model (ESM) via transfer learning and combines them with 3D structures of proteins predicted by AlphaFold2 through equivariant graph neural networks. Benchmarked on the CAFA3 test dataset and a new test dataset, TransFun outperforms several state-of-the-art methods, indicating the language model and 3D-equivariant graph neural networks are effective methods to leverage protein sequences and structures to improve protein function prediction. Combining TransFun predictions and sequence similarity-based predictions can further increase prediction accuracy.

Availability

The source code of TransFun is available at https://github.com/jianlin-cheng/TransFun.

Contact

chengji@missouri.edu.

Elucidating the functional roles of prokaryotic proteins using big data and artificial intelligence

Annotating protein sequences according to their biological functions is one of the key steps in understanding microbial diversity, metabolic potentials, and evolutionary histories. However, even in the best-studied prokaryotic genomes, not all proteins can be characterized by classical in vivo, in vitro, and/or in silico methods-a challenge rapidly growing alongside the advent of next-generation sequencing technologies and their enormous extension of 'omics' data in public databases. These so-called hypothetical proteins (HPs) represent a huge knowledge gap and hidden potential for biotechnological applications. Opportunities for leveraging the available 'Big Data' have recently proliferated with the use of artificial intelligence (AI). Here, we review the aims and methods of protein annotation and explain the different principles behind machine and deep learning algorithms including recent research examples, in order to assist both biologists wishing to apply AI tools in developing comprehensive genome annotations and computer scientists who want to contribute to this leading edge of biological research.

A novel hybrid CNN and BiGRU-Attention based deep learning model for protein function prediction

Proteins are the building blocks of all living things. Protein function must be ascertained if the molecular mechanism of life is to be understood. While CNN is good at capturing short-term relationships, GRU and LSTM can capture long-term dependencies. A hybrid approach that combines the complementary benefits of these deep-learning models motivates our work. Protein Language models, which use attention networks to gather meaningful data and build representations for proteins, have seen tremendous success in recent years processing the protein sequences. In this paper, we propose a hybrid CNN + BiGRU - Attention based model with protein language model embedding that effectively combines the output of CNN with the output of BiGRU-Attention for predicting protein functions. We evaluated the performance of our proposed hybrid model on human and yeast datasets. The proposed hybrid model improves the Fmax value over the state-of-the-art model SDN2GO for the cellular component prediction task by 1.9 %, for the molecular function prediction task by 3.8 % and for the biological process prediction task by 0.6 % for human dataset and for yeast dataset the cellular component prediction task by 2.4 %, for the molecular function prediction task by 5.2 % and for the biological process prediction task by 1.2 %.

PFP-GO: Integrating protein sequence, domain and protein-protein interaction information for protein function prediction using ranked GO terms

Protein function prediction is gradually emerging as an essential field in biological and computational studies. Though the latter has clinched a significant footprint, it has been observed that the application of computational information gathered from multiple sources has more significant influence than the one derived from a single source. Considering this fact, a methodology, PFP-GO, is proposed where heterogeneous sources like Protein Sequence, Protein Domain, and Protein-Protein Interaction Network have been processed separately for ranking each individual functional GO term. Based on this ranking, GO terms are propagated to the target proteins. While Protein sequence enriches the sequence-based information, Protein Domain and Protein-Protein Interaction Networks embed structural/functional and topological based information, respectively, during the phase of GO ranking. Performance analysis of PFP-GO is also based on Precision, Recall, and F-Score. The same was found to perform reasonably better when compared to the other existing state-of-art. PFP-GO has achieved an overall Precision, Recall, and F-Score of 0.67, 0.58, and 0.62, respectively. Furthermore, we check some of the top-ranked GO terms predicted by PFP-GO through multilayer network propagation that affect the 3D structure of the genome. The complete source code of PFP-GO is freely available at https://sites.google.com/view/pfp-go/.

Orthology-based analysis helps map evolutionary diversification and predict substrate class use of BAHD acyltransferases

Large enzyme families catalyze metabolic diversification by virtue of their ability to use diverse chemical scaffolds. How enzyme families attain such functional diversity is not clear. Furthermore, duplication and promiscuity in such enzyme families limits their functional prediction, which has produced a burgeoning set of incompletely annotated genes in plant genomes. Here, we address these challenges using BAHD acyltransferases as a model. This fast-evolving family expanded drastically in land plants, increasing from one to five copies in algae to approximately 100 copies in diploid angiosperm genomes. Compilation of >160 published activities helped visualize the chemical space occupied by this family and define eight different classes based on structural similarities between acceptor substrates. Using orthologous groups (OGs) across 52 sequenced plant genomes, we developed a method to predict BAHD acceptor substrate class utilization as well as origins of individual BAHD OGs in plant evolution. This method was validated using six novel and 28 previously characterized enzymes and helped improve putative substrate class predictions for BAHDs in the tomato genome. Our results also revealed that while cuticular wax and lignin biosynthetic activities were more ancient, anthocyanin acylation activity was fixed in BAHDs later near the origin of angiosperms. The OG-based analysis enabled identification of signature motifs in anthocyanin-acylating BAHDs, whose importance was validated via molecular dynamic simulations, site-directed mutagenesis and kinetic assays. Our results not only describe how BAHDs contributed to evolution of multiple chemical phenotypes in the plant world but also propose a biocuration-enabled approach for improved functional annotation of plant enzyme families.

The field of protein function prediction as viewed by different domain scientists

Motivation

Experimental biologists, biocurators, and computational biologists all play a role in characterizing a protein's function. The discovery of protein function in the laboratory by experimental scientists is the foundation of our knowledge about proteins. Experimental findings are compiled in knowledgebases by biocurators to provide standardized, readily accessible, and computationally amenable information. Computational biologists train their methods using these data to predict protein function and guide subsequent experiments. To understand the state of affairs in this ecosystem, centered here around protein function prediction, we surveyed scientists from these three constituent communities.

Results

We show that the three communities have common but also idiosyncratic perspectives on the field. Most strikingly, experimentalists rarely use state-of-the-art prediction software, but when presented with predictions, report many to be surprising and useful. Ontologies appear to be highly valued by biocurators, less so by experimentalists and computational biologists, yet controlled vocabularies bridge the communities and simplify the prediction task. Additionally, many software tools are not readily accessible and the predictions presented to the users can be broad and uninformative. We conclude that to meet both the social and technical challenges in the field, a more productive and meaningful interaction between members of the core communities is necessary.

Availability and implementation

Data cannot be shared for ethical/privacy reasons.

Supplementary information

Supplementary data are available at Bioinformatics Advances online.

Hierarchical deep learning for predicting GO annotations by integrating protein knowledge

MOTIVATION

Experimental testing and manual curation are the most precise ways for assigning Gene Ontology (GO) terms describing protein functions. However, they are expensive, time-consuming and cannot cope with the exponential growth of data generated by high-throughput sequencing methods. Hence, researchers need reliable computational systems to help fill the gap with automatic function prediction. The results of the last Critical Assessment of Function Annotation challenge revealed that GO-terms prediction remains a very challenging task. Recent developments on deep learning are significantly breaking out the frontiers leading to new knowledge in protein research thanks to the integration of data from multiple sources. However, deep models hitherto developed for functional prediction are mainly focused on sequence data and have not achieved breakthrough performances yet.

RESULTS

We propose DeeProtGO, a novel deep-learning model for predicting GO annotations by integrating protein knowledge. DeeProtGO was trained for solving 18 different prediction problems, defined by the three GO sub-ontologies, the type of proteins, and the taxonomic kingdom. Our experiments reported higher prediction quality when more protein knowledge is integrated. We also benchmarked DeeProtGO against state-of-the-art methods on public datasets, and showed it can effectively improve the prediction of GO annotations.

AVAILABILITY AND IMPLEMENTATION

DeeProtGO and a case of use are available at https://github.com/gamerino/DeeProtGO.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Identifying Functions of Proteins in Mice With Functional Embedding Features

In current biology, exploring the biological functions of proteins is important. Given the large number of proteins in some organisms, exploring their functions one by one through traditional experiments is impossible. Therefore, developing quick and reliable methods for identifying protein functions is necessary. Considerable accumulation of protein knowledge and recent developments on computer science provide an alternative way to complete this task, that is, designing computational methods. Several efforts have been made in this field. Most previous methods have adopted the protein sequence features or directly used the linkage from a protein–protein interaction (PPI) network. In this study, we proposed some novel multi-label classifiers, which adopted new embedding features to represent proteins. These features were derived from functional domains and a PPI network via word embedding and network embedding, respectively. The minimum redundancy maximum relevance method was used to assess the features, generating a feature list. Incremental feature selection, incorporating RAndom k-labELsets to construct multi-label classifiers, used such list to construct two optimum classifiers, corresponding to two key measurements: accuracy and exact match. These two classifiers had good performance, and they were superior to classifiers that used features extracted by traditional methods.

ContactPFP: Protein Function Prediction Using Predicted Contact Information

Computational function prediction is one of the most important problems in bioinformatics as elucidating the function of genes is a central task in molecular biology and genomics. Most of the existing function prediction methods use protein sequences as the primary source of input information because the sequence is the most available information for query proteins. There are attempts to consider other attributes of query proteins. Among these attributes, the three-dimensional (3D) structure of proteins is known to be very useful in identifying the evolutionary relationship of proteins, from which functional similarity can be inferred. Here, we report a novel protein function prediction method, ContactPFP, which uses predicted residue-residue contact maps as input structural features of query proteins. Although 3D structure information is known to be useful, it has not been routinely used in function prediction because the 3D structure is not experimentally determined for many proteins. In ContactPFP, we overcome this limitation by using residue-residue contact prediction, which has become increasingly accurate due to rapid development in the protein structure prediction field. ContactPFP takes a query protein sequence as input and uses predicted residue-residue contact as a proxy for the 3D protein structure. To characterize how predicted contacts contribute to function prediction accuracy, we compared the performance of ContactPFP with several well-established sequence-based function prediction methods. The comparative study revealed the advantages and weaknesses of ContactPFP compared to contemporary sequence-based methods. There were many cases where it showed higher prediction accuracy. We examined factors that affected the accuracy of ContactPFP using several illustrative cases that highlight the strength of our method.

Conditional generative modeling for de novo protein design with hierarchical functions

Motivation

Protein design has become increasingly important for medical and biotechnological applications. Because of the complex mechanisms underlying protein formation, the creation of a novel protein requires tedious and time-consuming computational or experimental protocols. At the same time, machine learning has enabled the solving of complex problems by leveraging large amounts of available data, more recently with great improvements on the domain of generative modeling. Yet, generative models have mainly been applied to specific sub-problems of protein design.

Results

Here, we approach the problem of general-purpose protein design conditioned on functional labels of the hierarchical Gene Ontology. Since a canonical way to evaluate generative models in this domain is missing, we devise an evaluation scheme of several biologically and statistically inspired metrics. We then develop the conditional generative adversarial network ProteoGAN and show that it outperforms several classic and more recent deep-learning baselines for protein sequence generation. We further give insights into the model by analyzing hyperparameters and ablation baselines. Lastly, we hypothesize that a functionally conditional model could generate proteins with novel functions by combining labels and provide first steps into this direction of research.

Availability and implementation

The code and data underlying this article are available on GitHub at https://github.com/timkucera/proteogan, and can be accessed with doi:10.5281/zenodo.6591379.

Supplementary information

Supplemental data are available at Bioinformatics online.

Diminished Auxin Signaling Triggers Cellular Reprogramming by Inducing a Regeneration Factor in the Liverwort Marchantia polymorpha

Regeneration in land plants is accompanied by the establishment of new stem cells, which often involves reactivation of the cell division potential in differentiated cells. The phytohormone auxin plays pivotal roles in this process. In bryophytes, regeneration is enhanced by the removal of the apex and repressed by exogenously applied auxin, which has long been proposed as a form of apical dominance. However, the molecular basis behind these observations remains unexplored. Here, we demonstrate that in the liverwort Marchantia polymorpha, the level of endogenous auxin is transiently decreased in the cut surface of decapitated explants, and identify by transcriptome analysis a key transcription factor gene, LOW-AUXIN RESPONSIVE (MpLAXR), which is induced upon auxin reduction. Loss of MpLAXR function resulted in delayed cell cycle reactivation, and transient expression of MpLAXR was sufficient to overcome the inhibition of regeneration by exogenously applied auxin. Furthermore, ectopic expression of MpLAXR caused cell proliferation in normally quiescent tissues. Together, these data indicate that decapitation causes a reduction of auxin level at the cut surface, where, in response, MpLAXR is up-regulated to trigger cellular reprogramming. MpLAXR is an ortholog of Arabidopsis ENHANCER OF SHOOT REGENERATION 1/DORNRÖSCHEN, which has dual functions as a shoot regeneration factor and a regulator of axillary meristem initiation, the latter of which requires a low auxin level. Thus, our findings provide insights into stem cell regulation as well as apical dominance establishment in land plants.

Deep neural learning based protein function prediction

It is vital for the annotation of uncharacterized proteins by protein function prediction. At present, Deep Neural Network based protein function prediction is mainly carried out for dataset of small scale proteins or Gene Ontology, and usually explore the relationships between single protein feature and function tags. The practical methods for large-scale multi-features protein prediction still need to be studied in depth. This paper proposes a DNN based protein function prediction approach IGP-DNN. This method uses Grasshopper Optimization Algorithm (GOA) and Intuitionistic Fuzzy c-Means clustering (IFCM) based protein function modules extracting algorithm to extract the features of protein modules, utilizing Kernel Principal Component Analysis (KPCA) method to reduce the dimensionality of the protein attribute information, and integrating module features and attribute features. Inputting integrated data into DNN through multiple hidden layers to classify proteins and predict protein functions. In the experiments, the F-measure value of IGP-DNN on the DIP dataset reaches 0.4436, which shows better performance.

PANNZER-A practical tool for protein function prediction

The facility of next-generation sequencing has led to an explosion of gene catalogs for novel genomes, transcriptomes and metagenomes, which are functionally uncharacterized. Computational inference has emerged as a necessary substitute for first-hand experimental evidence. PANNZER (Protein ANNotation with Z-scoRE) is a high-throughput functional annotation web server that stands out among similar publically accessible web servers in supporting submission of up to 100,000 protein sequences at once and providing both Gene Ontology (GO) annotations and free text description predictions. Here, we demonstrate the use of PANNZER and discuss future plans and challenges. We present two case studies to illustrate problems related to data quality and method evaluation. Some commonly used evaluation metrics and evaluation datasets promote methods that favor unspecific and broad functional classes over more informative and specific classes. We argue that this can bias the development of automated function prediction methods. The PANNZER web server and source code are available at http://ekhidna2.biocenter.helsinki.fi/sanspanz/.

Accurate protein function prediction via graph attention networks with predicted structure information

Experimental protein function annotation does not scale with the fast-growing sequence databases. Only a tiny fraction (<0.1%) of protein sequences has experimentally determined functional annotations. Computational methods may predict protein function very quickly, but their accuracy is not very satisfactory. Based upon recent breakthroughs in protein structure prediction and protein language models, we develop GAT-GO, a graph attention network (GAT) method that may substantially improve protein function prediction by leveraging predicted structure information and protein sequence embedding. Our experimental results show that GAT-GO greatly outperforms the latest sequence- and structure-based deep learning methods. On the PDB-mmseqs testset where the train and test proteins share <15% sequence identity, our GAT-GO yields Fmax (maximum F-score) 0.508, 0.416, 0.501, and area under the precision-recall curve (AUPRC) 0.427, 0.253, 0.411 for the MFO, BPO, CCO ontology domains, respectively, much better than the homology-based method BLAST (Fmax 0.117, 0.121, 0.207 and AUPRC 0.120, 0.120, 0.163) that does not use any structure information. On the PDB-cdhit testset where the training and test proteins are more similar, although using predicted structure information, our GAT-GO obtains Fmax 0.637, 0.501, 0.542 for the MFO, BPO, CCO ontology domains, respectively, and AUPRC 0.662, 0.384, 0.481, significantly exceeding the just-published method DeepFRI that uses experimental structures, which has Fmax 0.542, 0.425, 0.424 and AUPRC only 0.313, 0.159, 0.193.

Protein function prediction using functional inter-relationship

With the growth of high throughput sequencing techniques, the generation of protein sequences has become fast and cheap, leading to a huge increase in the number of known proteins. However, it is challenging to identify the functions being performed by these newly discovered proteins. Machine learning techniques have improved traditional methods' efficiency by suggesting relevant functions but fails to perform well when the number of functions to be predicted becomes large. In this work, we propose a machine learning-based approach to predict huge set of protein functions that use the inter-relationships between functions to improve the model's predictability. These inter-relationships of functions is used to reduce the redundancy caused by highly correlated functions. The proposed model is trained on the reduced set of non-redundant functions hindering the ambiguity caused due to inter-related functions. Here, we use two statistical approaches 1) Pearson's correlation coefficient 2) Jaccard similarity coefficient, as a measure of correlation to remove redundant functions. To have a fair evaluation of the proposed model, we recreate our original function set by inverse transforming the reduced set using the two proposed approaches: Direct mapping and Ensemble approach. The model is tested using different feature sets and function sets of biological processes and molecular functions to get promising results on DeepGO and CAFA3 dataset. The proposed model is able to predict specific functions for the test data which were unpredictable by other compared methods. The experimental models, code and other relevant data are available at https://github.com/richadhanuka/PFP-using-Functional-interrelationship.

A Deep Learning Framework for Gene Ontology Annotations With Sequence- and Network-Based Information

Knowledge of protein functions plays an important role in biology and medicine. With the rapid development of high-throughput technologies, a huge number of proteins have been discovered. However, there are a great number of proteins without functional annotations. A protein usually has multiple functions and some functions or biological processes require interactions of a plurality of proteins. Additionally, Gene Ontology provides a useful classification for protein functions and contains more than 40,000 terms. We propose a deep learning framework called DeepGOA to predict protein functions with protein sequences and protein-protein interaction (PPI) networks. For protein sequences, we extract two types of information: sequence semantic information and subsequence-based features. We use the word2vec technique to numerically represent protein sequences, and utilize a Bi-directional Long and Short Time Memory (Bi-LSTM) and multi-scale convolutional neural network (multi-scale CNN) to obtain the global and local semantic features of protein sequences, respectively. Additionally, we use the InterPro tool to scan protein sequences for extracting subsequence-based information, such as domains and motifs. Then, the information is plugged into a neural network to generate high-quality features. For the PPI network, the Deepwalk algorithm is applied to generate its embedding information of PPI. Then the two types of features are concatenated together to predict protein functions. To evaluate the performance of DeepGOA, several different evaluation methods and metrics are utilized. The experimental results show that DeepGOA outperforms DeepGO and BLAST.

Protein function prediction with gene ontology: from traditional to deep learning models

Protein function prediction is a crucial part of genome annotation. Prediction methods have recently witnessed rapid development, owing to the emergence of high-throughput sequencing technologies. Among the available databases for identifying protein function terms, Gene Ontology (GO) is an important resource that describes the functional properties of proteins. Researchers are employing various approaches to efficiently predict the GO terms. Meanwhile, deep learning, a fast-evolving discipline in data-driven approach, exhibits impressive potential with respect to assigning GO terms to amino acid sequences. Herein, we reviewed the currently available computational GO annotation methods for proteins, ranging from conventional to deep learning approach. Further, we selected some suitable predictors from among the reviewed tools and conducted a mini comparison of their performance using a worldwide challenge dataset. Finally, we discussed the remaining major challenges in the field, and emphasized the future directions for protein function prediction with GO.

ARG-SHINE: improve antibiotic resistance class prediction by integrating sequence homology, functional information and deep convolutional neural network

Antibiotic resistance in bacteria limits the effect of corresponding antibiotics, and the classification of antibiotic resistance genes (ARGs) is important for the treatment of bacterial infections and for understanding the dynamics of microbial communities. Although several methods have been developed to classify ARGs, none of them work well when the ARGs diverge from those in the reference ARG databases. We develop a novel method, ARG-SHINE, for ARG classification. ARG-SHINE utilizes state-of-the-art learning to rank machine learning approach to ensemble three component methods with different features, including sequence homology, protein domain/family/motif and raw amino acid sequences for the deep convolutional neural network. Compared with other methods, ARG-SHINE achieves better performance on two benchmark datasets in terms of accuracy, macro-average f1-score and weighted-average f1-score. ARG-SHINE is used to classify newly discovered ARGs through functional screening and achieves high prediction accuracy. ARG-SHINE is freely available at https://github.com/ziyewang/ARG_SHINE.

NetGO 2.0: improving large-scale protein function prediction with massive sequence, text, domain, family and network information

With the explosive growth of protein sequences, large-scale automated protein function prediction (AFP) is becoming challenging. A protein is usually associated with dozens of gene ontology (GO) terms. Therefore, AFP is regarded as a problem of large-scale multi-label classification. Under the learning to rank (LTR) framework, our previous NetGO tool integrated massive networks and multi-type information about protein sequences to achieve good performance by dealing with all possible GO terms (>44 000). In this work, we propose the updated version as NetGO 2.0, which further improves the performance of large-scale AFP. NetGO 2.0 also incorporates literature information by logistic regression and deep sequence information by recurrent neural network (RNN) into the framework. We generate datasets following the critical assessment of functional annotation (CAFA) protocol. Experiment results show that NetGO 2.0 outperformed NetGO significantly in biological process ontology (BPO) and cellular component ontology (CCO). In particular, NetGO 2.0 achieved a 12.6% improvement over NetGO in terms of area under precision-recall curve (AUPR) in BPO and around 2.6% in terms of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\mathbf {F_{max}}$\end{document} in CCO. These results demonstrate the benefits of incorporating text and deep sequence information for the functional annotation of BPO and CCO. The NetGO 2.0 web server is freely available at http://issubmission.sjtu.edu.cn/ng2/.

DeepGraphGO: graph neural network for large-scale, multispecies protein function prediction

Motivation

Automated function prediction (AFP) of proteins is a large-scale multi-label classification problem. Two limitations of most network-based methods for AFP are (i) a single model must be trained for each species and (ii) protein sequence information is totally ignored. These limitations cause weaker performance than sequence-based methods. Thus, the challenge is how to develop a powerful network-based method for AFP to overcome these limitations.

Results

We propose DeepGraphGO, an end-to-end, multispecies graph neural network-based method for AFP, which makes the most of both protein sequence and high-order protein network information. Our multispecies strategy allows one single model to be trained for all species, indicating a larger number of training samples than existing methods. Extensive experiments with a large-scale dataset show that DeepGraphGO outperforms a number of competing state-of-the-art methods significantly, including DeepGOPlus and three representative network-based methods: GeneMANIA, deepNF and clusDCA. We further confirm the effectiveness of our multispecies strategy and the advantage of DeepGraphGO over so-called difficult proteins. Finally, we integrate DeepGraphGO into the state-of-the-art ensemble method, NetGO, as a component and achieve a further performance improvement.

Availability and implementation

https://github.com/yourh/DeepGraphGO.

Supplementary information

Supplementary data are available at Bioinformatics online.

DeepGOWeb: fast and accurate protein function prediction on the (Semantic) Web

Understanding the functions of proteins is crucial to understand biological processes on a molecular level. Many more protein sequences are available than can be investigated experimentally. DeepGOPlus is a protein function prediction method based on deep learning and sequence similarity. DeepGOWeb makes the prediction model available through a website, an API, and through the SPARQL query language for interoperability with databases that rely on Semantic Web technologies. DeepGOWeb provides accurate and fast predictions and ensures that predicted functions are consistent with the Gene Ontology; it can provide predictions for any protein and any function in Gene Ontology. DeepGOWeb is freely available at https://deepgo.cbrc.kaust.edu.sa/.

IsoResolve: predicting splice isoform functions by integrating gene and isoform-level features with domain adaptation

MOTIVATION

High resolution annotation of gene functions is a central goal in functional genomics. A single gene may produce multiple isoforms with different functions through alternative splicing. Conventional approaches, however, consider a gene as a single entity without differentiating these functionally different isoforms. Towards understanding gene functions at higher resolution, recent efforts have focused on predicting the functions of isoforms. However, the performance of existing methods is far from satisfactory mainly because of the lack of isoform-level functional annotation.

RESULTS

We present IsoResolve, a novel approach for isoform function prediction, which leverages the information from gene function prediction models with domain adaptation (DA). IsoResolve treats gene-level and isoform-level features as source and target domains, respectively. It uses DA to project the two domains into a latent variable space in such a way that the latent variables from the two domains have similar distribution, which enables the gene domain information to be leveraged for isoform function prediction. We systematically evaluated the performance of IsoResolve in predicting functions. Compared with five state-of-the-art methods, IsoResolve achieved significantly better performance. IsoResolve was further validated by case studies of genes with isoform-level functional annotation.

AVAILABILITY AND IMPLEMENTATION

IsoResolve is freely available at https://github.com/genemine/IsoResolve.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

TALE: Transformer-based protein function Annotation with joint sequence-Label Embedding

MOTIVATION

Facing the increasing gap between high-throughput sequence data and limited functional insights, computational protein function annotation provides a high-throughput alternative to experimental approaches. However, current methods can have limited applicability while relying on protein data besides sequences, or lack generalizability to novel sequences, species and functions.

RESULTS

To overcome aforementioned barriers in applicability and generalizability, we propose a novel deep learning model using only sequence information for proteins, named Transformer-based protein function Annotation through joint sequence-Label Embedding (TALE). For generalizability to novel sequences we use self-attention-based transformers to capture global patterns in sequences. For generalizability to unseen or rarely seen functions (tail labels), we embed protein function labels (hierarchical GO terms on directed graphs) together with inputs/features (1D sequences) in a joint latent space. Combining TALE and a sequence similarity-based method, TALE+ outperformed competing methods when only sequence input is available. It even outperformed a state-of-the-art method using network information besides sequence, in two of the three gene ontologies. Furthermore, TALE and TALE+ showed superior generalizability to proteins of low similarity, new species, or rarely annotated functions compared to training data, revealing deep insights into the protein sequence-function relationship. Ablation studies elucidated contributions of algorithmic components toward the accuracy and the generalizability; and a GO term-centric analysis was also provided.

AVAILABILITY AND IMPLEMENTATION

The data, source codes and models are available at https://github.com/Shen-Lab/TALE.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

NetQuilt: Deep Multispecies Network-based Protein Function Prediction using Homology-informed Network Similarity

Motivation

Transferring knowledge between species is challenging: different species contain distinct proteomes and cellular architectures, which cause their proteins to carry out different functions via different interaction networks. Many approaches to protein functional annotation use sequence similarity to transfer knowledge between species. These approaches cannot produce accurate predictions for proteins without homologues of known function, as many functions require cellular context for meaningful prediction. To supply this context, network-based methods use protein-protein interaction (PPI) networks as a source of information for inferring protein function and have demonstrated promising results in function prediction. However, most of these methods are tied to a network for a single species, and many species lack biological networks.

Results

In this work, we integrate sequence and network information across multiple species by computing IsoRank similarity scores to create a meta-network profile of the proteins of multiple species. We use this integrated multispecies meta-network as input to train a maxout neural network with Gene Ontology terms as target labels. Our multispecies approach takes advantage of more training examples, and consequently leads to significant improvements in function prediction performance compared to two network-based methods, a deep learning sequence-based method and the BLAST annotation method used in the Critial Assessment of Functional Annotation. We are able to demonstrate that our approach performs well even in cases where a species has no network information available: when an organism’s PPI network is left out we can use our multi-species method to make predictions for the left-out organism with good performance.

Availability and implementation

The code is freely available at https://github.com/nowittynamesleft/NetQuilt. The data, including sequences, PPI networks and GO annotations are available at https://string-db.org/.

Supplementary information

Supplementary data are available at Bioinformatics online.