DeepGOPlus: Improved protein function prediction from sequence

Kulmanov, Maxat; Hoehndorf, Robert

doi:10.1101/615260

Cited by 30 publications

(52 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, it is designed to predict phenotypes which arise from a complete loss of function of a gene. Together with function prediction methods such as DeepGOPlus (Kulmanov and Hoehndorf, 2019), DeepPheno can, in principle, predict phenotype associations for protein-coding genes using only the protein's amino acid sequence. However, DeepGOPlus was trained on experimentally annotated sequences of many organisms, including several animal model organisms.…”

Section: Discussionmentioning

confidence: 99%

“…We filter experimental annotations using the evidence codes EXP, IDA, IPI, IMP, IGI, IEP, TAS, IC, HTP, HDA, HMP, HGI, HEP. For the third dataset, we use GO functions predicted by DeepGOPlus (Kulmanov and Hoehndorf, 2019).…”

Section: Training and Testing Datasetmentioning

confidence: 99%

“…We use OMIM (Hamosh et al, 2005) diseases and their phenotype annotations from the HPO (Köhler et al, 2017) to learn associations between sets of GO functional annotations and sets of HPO phenotype annotations. Since experimental annotations are not available for all genes, we also utilize the sequence-based function prediction method DeepGOPlus (Kulmanov and Hoehndorf, 2019), which uses information of function annotations in many organisms, to fill this gap and predict the functions of gene products. Consequently, DeepPheno can predict phenotypes for all protein-coding genes with an available sequence.…”

Section: Deeppheno Modelmentioning

confidence: 99%

“…One limitation of predicting phenotypes from functions is that not all genes have experimental functional annotations. We overcome this limitation by using the function prediction method DeepGOPlus (Kulmanov and Hoehndorf, 2019) which has been trained on a large number of proteins and can generate accurate functional annotations using only protein sequence. We evaluate our method using the latest phenotype annotations from the HPO (Köhler et al, 2017) and using the evaluation dataset from the Computational Assessment of Function Annotation (CAFA) challenge , and we compare our results with the top performing methods in CAFA 2 (Jiang et al, 2016) and one recent phenotype prediction method, HPO2GO (Dogan, 2018).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

DeepPheno: Predicting single gene loss-of-function phenotypes using an ontology-aware hierarchical classifier

Kulmanov

Hoehndorf

2019

Preprint

Self Cite

View full text Add to dashboard Cite

Motivation:Predicting the phenotypes resulting from molecular perturbations is one of the key challenges in genetics. Both forward and reverse genetic screen are employed to identify the molecular mechanisms underlying phenotypes and disease, and these resulted in a large number of genotype-phenotype association being available for humans and model organisms. Combined with recent advances in machine learning, it may now be possible to predict human phenotypes resulting from particular molecular aberrations. Results: We developed DeepPheno, a method for predicting the phenotypes resulting from complete loss-of-function in single genes. DeepPheno uses the functional annotations with gene products to predict the phenotypes resulting from a loss-of-function; additionally, we employ a two-step procedure in which we predict these functions first and then predict phenotypes. This allows us to predict phenotypes associated with any known protein-coding gene. We evaluate our approach using evaluation metrics established by the CAFA challenge and compare with top performing CAFA 2 methods. Our method achieves an Fmax of 0.46 which is a significant improvement over state-of-the-art Fmax of 0.36. Furthermore, we show that predictions generated by DeepPheno are applicable to predicting gene-disease associations based on comparing phenotypes, and 60% of predictions made by DeepPheno interact with a gene that is already associated with the predicted phenotype. Availability: https://

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Training and Testing Datasetmentioning

confidence: 99%

Section: Deeppheno Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

DeepPheno: Predicting single gene loss-of-function phenotypes using an ontology-aware hierarchical classifier

Kulmanov

Hoehndorf

2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Only proteins with at least three experimentally validated GO terms are kept, in an attempt to reduce an overestimation of false positive predictions as described by Dessimoz et al [6]. Initial predictions are done using Argot2.5 [7], InterProScan 5.21-60 [8], and DeepGOPlus [9], using only databases from before 2017 to ensure blind predictions. Half of the proteins are used to train an SVM model with CrowdGO_training, and half are used to predict GO terms using CrowdGO.…”

Section: Crowdgo Demonstrationmentioning

confidence: 99%

CrowdGO: machine learning and semantic similarity guided consensus Gene Ontology annotation

Reijnders

Waterhouse

2019

Preprint

View full text Add to dashboard Cite

Motivation:Protein function prediction tools vary widely in their methodologies, resulting in different sets of GO terms being correctly predicted. Ideally, multiple tools are combined to achieve a higher recall of GO terms while increasing precision.Results: CrowdGO combines input predictions from any number of tools and combines them based on the Gene Ontology Directed Acyclic Graph. Using each GO terms information content, the semantic similarity between GO predictions of different tools, and a Support Vector Machine model, it achieves improved precision and recall compared to each of the tools separately ( Figure 1).Availability: CrowdGO can be found at https://gitlab.com/mreijnders/CrowdGO

show abstract

Genome wide identification and expression analysis of pepper C2H2 zinc finger transcription factors in response to anthracnose pathogen Colletotrichum truncatum

et al. 2021

View full text Add to dashboard Cite

Although, the C 2 H 2 zinc finger (ZF) family of plant transcription factors have been implicated in multiple biological processes, they are yet to be characterized in the economically important chilli pepper (Capsicum annuum). In this study, a total of 79 C 2 H 2 ZF genes were identified in the pepper genome. Phylogenetic analysis categorized the pepper C 2 H 2 ZF (CaZF) members into five subfamilies each with unique conserved domains and functions. Genomic organization revealed that CaZF genes have variable number of introns consistent with the characteristics defined by the evolutionary analysis. Segmental duplication-based purifying selection contributed to the expansion of CaZF genes in pepper. Additionally, 11 CaZF genes were identified as targets for 38 miRNAs indicating their role in post-transcriptional silencing-mediated genetic regulation. Gene expression analysis revealed that 18 CaZF genes were differentially expressed post-infection with the anthrocnose pathogen Colletotrichum truncatum, uncovering their potential function in pepper response to biotic stresses. Moreover, CaZFs were significantly induced post-treatment with methyl jasmonate and ethylene indicating their role in defense signaling. Notably, the MeJA responsive cis-elements were detected in the promoter regions of majority of CaZF genes, suggesting that CaZFs may be implicated in defense-responsive signal cross talking. Additionally, 18 CaZF genes were differentially expressed under drought and heat treatment, indicating their involvement in plant response to abiotic stresses. Overall, a comprehensive analysis of CaZF gene family in pepper provided significant insights into the understanding of C 2 H 2 ZFmediated stress regulation network, which would benefit the genetic improvement of pepper and other allied plants. KeywordsCapsicum annuum • C 2 H 2 zinc finger proteins • Genome wide analysis • Pepper anthracnose • Colletotrichum truncatum

show abstract

DeepGOPlus: Improved protein function prediction from sequence

Cited by 30 publications

References 12 publications

DeepPheno: Predicting single gene loss-of-function phenotypes using an ontology-aware hierarchical classifier

DeepPheno: Predicting single gene loss-of-function phenotypes using an ontology-aware hierarchical classifier

CrowdGO: machine learning and semantic similarity guided consensus Gene Ontology annotation

Genome wide identification and expression analysis of pepper C2H2 zinc finger transcription factors in response to anthracnose pathogen Colletotrichum truncatum

Contact Info

Product

Resources

About