scholarly journals Embeddings from deep learning transfer GO annotations beyond homology

2020 ◽  
Author(s):  
Maria Littmann ◽  
Michael Heinzinger ◽  
Christian Dallago ◽  
Tobias Olenyi ◽  
Burkhard Rost
Keyword(s):  
Protein Function ◽  
Protein Sequences ◽  
Language Models ◽  
Evolutionary Information ◽  
Pairwise Sequence Identity ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Sequence Identity ◽  
Experimental Function ◽  
Go Terms

AbstractKnowing 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.

scholarly journals Embeddings from deep learning transfer GO annotations beyond homology

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Maria Littmann ◽  
Michael Heinzinger ◽  
Christian Dallago ◽  
Tobias Olenyi ◽  
Burkhard Rost
Keyword(s):  
Protein Function ◽  
Protein Sequences ◽  
Language Models ◽  
Evolutionary Information ◽  
Pairwise Sequence Identity ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Sequence Identity ◽  
Experimental Function ◽  
Go Terms

AbstractKnowing 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.

scholarly journals INGA 2.0: improving protein function prediction for the dark proteome

2019 ◽  
Vol 47 (W1) ◽  
pp. W373-W378 ◽  
Author(s):  
Damiano Piovesan ◽  
Silvio C E Tosatto
Keyword(s):  
Protein Function ◽  
Current Knowledge ◽  
Protein Function Prediction ◽  
Domain Architecture ◽  
Function Prediction ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Additional Input ◽  
Consensus Prediction ◽  
Go Terms

Abstract Our current knowledge of complex biological systems is stored in a computable form through the Gene Ontology (GO) which provides a comprehensive description of genes function. Prediction of GO terms from the sequence remains, however, a challenging task, which is particularly critical for novel genomes. Here we present INGA 2.0, a new version of the INGA software for protein function prediction. INGA exploits homology, domain architecture, interaction networks and information from the ‘dark proteome’, like transmembrane and intrinsically disordered regions, to generate a consensus prediction. INGA was ranked in the top ten methods on both CAFA2 and CAFA3 blind tests. The new algorithm can process entire genomes in a few hours or even less when additional input files are provided. The new interface provides a better user experience by integrating filters and widgets to explore the graph structure of the predicted terms. The INGA web server, databases and benchmarking are available from URL: https://inga.bio.unipd.it/.

scholarly journals Intrinsically disordered regions are abundant in simplexvirus proteomes and display signatures of positive selection

2020 ◽  
Vol 6 (1) ◽  
Author(s):  
Alessandra Mozzi ◽  
Diego Forni ◽  
Rachele Cagliani ◽  
Mario Clerici ◽  
Uberto Pozzoli ◽  
...  
Keyword(s):  
Positive Selection ◽  
Protein Function ◽  
Three Dimensional ◽  
New Host ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Human Proteins ◽  
Simplex Virus ◽  
Herpes Simplex Virus 2 ◽  
Disordered Regions

Abstract Whereas the majority of herpesviruses co-speciated with their mammalian hosts, human herpes simplex virus 2 (HSV-2, genus Simplexvirus) most likely originated from the cross-species transmission of chimpanzee herpesvirus 1 to an ancestor of modern humans. We exploited the peculiar evolutionary history of HSV-2 to investigate the selective events that drove herpesvirus adaptation to a new host. We show that HSV-2 intrinsically disordered regions (IDRs)—that is, protein domains that do not adopt compact three-dimensional structures—are strongly enriched in positive selection signals. Analysis of viral proteomes indicated that a significantly higher portion of simplexvirus proteins is disordered compared with the proteins of other human herpesviruses. IDR abundance in simplexvirus proteomes was not a consequence of the base composition of their genomes (high G + C content). Conversely, protein function determines the IDR fraction, which is significantly higher in viral proteins that interact with human factors. We also found that the average extent of disorder in herpesvirus proteins tends to parallel that of their human interactors. These data suggest that viruses that interact with fast-evolving, disordered human proteins, in turn, evolve disordered viral interactors poised for innovation. We propose that the high IDR fraction present in simplexvirus proteomes contributes to their wider host range compared with other herpesviruses.

scholarly journals Structural propensity database of proteins

2017 ◽  
Author(s):  
Kamil Tamiola ◽  
Matthew M Heberling ◽  
Jan Domanski
Keyword(s):  
Structural Dynamics ◽  
Protein Function ◽  
Chemical Shifts ◽  
Protein Disorder ◽  
Computational Biophysics ◽  
Intrinsic Protein ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Intrinsic Protein Disorder ◽  
Structural Insights

AbstractAn overwhelming amount of experimental evidence suggests that elucidations of protein function, interactions, and pathology are incomplete without inclusion of intrinsic protein disorder and structural dynamics. Thus, to expand our understanding of intrinsic protein disorder, we have created a database of secondary structure (SS) propensities for proteins (dSPP) as a reference resource for experimental research and computational biophysics. The dSPP comprises SS propensities of 7,094 unrelated proteins, as gauged from NMR chemical shift measurements in solution and solid state. Here, we explain the concept of SS propensity and analyze dSPP entries of therapeutic relevance, α-synuclein, MOAG-4, and the ZIKA NS2B-NS3 complex to show: (1) how propensity mapping generates novel structural insights into intrinsically disordered regions of pathologically relevant proteins, (2) how computational biophysics tools can benefit from propensity mapping, and (3) how the residual disorder estimation based on NMR chemical shifts compares with sequence-based disorder predictors. This work demonstrates the benefit of propensity estimation as a method that reports both on protein structure, lability, and disorder.

scholarly journals Prediction of MoRFs based on sequence properties and convolutional neural networks

2021 ◽  
Vol 14 (1) ◽  
Author(s):  
Hao He ◽  
Yatong Zhou ◽  
Yue Chi ◽  
Jingfei He
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Intrinsically Disordered Proteins ◽  
Prediction Method ◽  
Disordered Proteins ◽  
Evolutionary Information ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Functional Regions ◽  
Interaction Partners

Abstract Background Intrinsically disordered proteins possess flexible 3-D structures, which makes them play an important role in a variety of biological functions. Molecular recognition features (MoRFs) act as an important type of functional regions, which are located within longer intrinsically disordered regions and undergo disorder-to-order transitions upon binding their interaction partners. Results We develop a method, MoRFCNN, to predict MoRFs based on sequence properties and convolutional neural networks (CNNs). The sequence properties contain structural and physicochemical properties which are used to describe the differences between MoRFs and non-MoRFs. Especially, to highlight the correlation between the target residue and adjacent residues, three windows are selected to preprocess the selected properties. After that, these calculated properties are combined into the feature matrix to predict MoRFs through the constructed CNN. Comparing with other existing methods, MoRFCNN obtains better performance. Conclusions MoRFCNN is a new individual MoRFs prediction method which just uses protein sequence properties without evolutionary information. The simulation results show that MoRFCNN is effective and competitive.

scholarly journals Learning protein constitutive motifs from sequence data

eLife ◽  
2019 ◽  
Vol 8 ◽  
Author(s):  
Jérôme Tubiana ◽  
Simona Cocco ◽  
Rémi Monasson
Keyword(s):  
Sequence Data ◽  
Protein Sequences ◽  
Sequence Information ◽  
Ligand Specificity ◽  
Protein Families ◽  
Restricted Boltzmann Machines ◽  
Intrinsically Disordered ◽  
Intrinsically Disordered Regions ◽  
Model Protein ◽  
Lattice Proteins

Statistical analysis of evolutionary-related protein sequences provides information about their structure, function, and history. We show that Restricted Boltzmann Machines (RBM), designed to learn complex high-dimensional data and their statistical features, can efficiently model protein families from sequence information. We here apply RBM to 20 protein families, and present detailed results for two short protein domains (Kunitz and WW), one long chaperone protein (Hsp70), and synthetic lattice proteins for benchmarking. The features inferred by the RBM are biologically interpretable: they are related to structure (residue-residue tertiary contacts, extended secondary motifs (α-helixes and β-sheets) and intrinsically disordered regions), to function (activity and ligand specificity), or to phylogenetic identity. In addition, we use RBM to design new protein sequences with putative properties by composing and 'turning up' or 'turning down' the different modes at will. Our work therefore shows that RBM are versatile and practical tools that can be used to unveil and exploit the genotype–phenotype relationship for protein families.

scholarly journals Modeling the language of life – Deep Learning Protein Sequences

2019 ◽  
Author(s):  
Michael Heinzinger ◽  
Ahmed Elnaggar ◽  
Yu Wang ◽  
Christian Dallago ◽  
Dmitrii Nechaev ◽  
...  
Keyword(s):  
Language Processing ◽  
Protein Function ◽  
Protein Sequences ◽  
Water Soluble ◽  
Evolutionary Information ◽  
Biophysical Properties ◽  
Data Set ◽  
Link Type ◽  
Single Protein ◽  
Sequence Databases

AbstractBackgroundOne common task in Computational Biology is the prediction of aspects of protein function and structure from their amino acid sequence. For 26 years, most state-of-the-art approaches toward this end have been marrying machine learning and evolutionary information. The retrieval of related proteins from ever growing sequence databases is becoming so time-consuming that the analysis of entire proteomes becomes challenging. On top, evolutionary information is less powerful for small families, e.g. for proteins from the Dark Proteome.ResultsWe introduce a novel way to represent protein sequences as continuous vectors (embeddings) by using the deep bi-directional model ELMo taken from natural language processing (NLP). The model has effectively captured the biophysical properties of protein sequences from unlabeled big data (UniRef50). After training, this knowledge is transferred to single protein sequences by predicting relevant sequence features. We refer to these new embeddings as SeqVec (Sequence-to-Vector) and demonstrate their effectiveness by training simple convolutional neural networks on existing data sets for two completely different prediction tasks. At the per-residue level, we significantly improved secondary structure (for NetSurfP-2.0 data set: Q3=79%±1, Q8=68%±1) and disorder predictions (MCC=0.59±0.03) over methods not using evolutionary information. At the per-protein level, we predicted subcellular localization in ten classes (for DeepLoc data set: Q10=68%±1) and distinguished membrane-bound from water-soluble proteins (Q2= 87%±1). All results built upon the embeddings gained from the new tool SeqVec neither explicitly nor implicitly using evolutionary information. Nevertheless, it improved over some methods using such information. Where the lightning-fast HHblits needed on average about two minutes to generate the evolutionary information for a target protein, SeqVec created the vector representation on average in 0.03 seconds.ConclusionWe have shown that transfer learning can be used to capture biochemical or biophysical properties of protein sequences from large unlabeled sequence databases. The effectiveness of the proposed approach was showcased for different prediction tasks using only single protein sequences. SeqVec embeddings enable predictions that outperform even some methods using evolutionary information. Thus, they prove to condense the underlying principles of protein sequences. This might be the first step towards competitive predictions based only on single protein sequences.AvailabilitySeqVec: https://github.com/mheinzinger/SeqVec Prediction server: https://embed.protein.properties

Sign in / Sign up

Export Citation Format

Share Document