MusiteDeep: a deep-learning based webserver for protein post-translational modification site prediction and visualization

Wang, Duolin; Liu, Dongpeng; Yuchi, Jiakang; He, Fei; Jiang, Yuyong; Cai, Siteng; Li, Jingyi; Xu, Dong

doi:10.1093/nar/gkaa275

Cited by 171 publications

(153 citation statements)

References 31 publications

Supporting

Mentioning

128

Contrasting

Order By: Relevance

“…To validate the performance of MultiLyGAN, the comparison of our models with the MusiteDeep [ 45 ] was performed. MusiteDeep, a deep-learning based predictor, provided identification for multiple PTMs, including 13 PTMs of which five are lysine-based modifications (Comparisons of number of enrolled proteins and modification sites were in Table 5 of Additional file: S3 ).…”

Section: Discussionmentioning

confidence: 99%

Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks

Yang

Wang

et al. 2021

BMC Bioinformatics

View full text Add to dashboard Cite

Background Protein post-translational modification (PTM) is a key issue to investigate the mechanism of protein’s function. With the rapid development of proteomics technology, a large amount of protein sequence data has been generated, which highlights the importance of the in-depth study and analysis of PTMs in proteins. Method We proposed a new multi-classification machine learning pipeline MultiLyGAN to identity seven types of lysine modified sites. Using eight different sequential and five structural construction methods, 1497 valid features were remained after the filtering by Pearson correlation coefficient. To solve the data imbalance problem, Conditional Generative Adversarial Network (CGAN) and Conditional Wasserstein Generative Adversarial Network (CWGAN), two influential deep generative methods were leveraged and compared to generate new samples for the types with fewer samples. Finally, random forest algorithm was utilized to predict seven categories. Results In the tenfold cross-validation, accuracy (Acc) and Matthews correlation coefficient (MCC) were 0.8589 and 0.8376, respectively. In the independent test, Acc and MCC were 0.8549 and 0.8330, respectively. The results indicated that CWGAN better solved the existing data imbalance and stabilized the training error. Alternatively, an accumulated feature importance analysis reported that CKSAAP, PWM and structural features were the three most important feature-encoding schemes. MultiLyGAN can be found at https://github.com/Lab-Xu/MultiLyGAN. Conclusions The CWGAN greatly improved the predictive performance in all experiments. Features derived from CKSAAP, PWM and structure schemes are the most informative and had the greatest contribution to the prediction of PTM.

show abstract

Section: Discussionmentioning

confidence: 99%

Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks

Yang

Wang

et al. 2021

BMC Bioinformatics

View full text Add to dashboard Cite

show abstract

“…The overall models were next subjected to repeated rounds of local and global optimization in NAMD [ 56 ] using CHARMM36 force field, followed in an iterative manner till convergence by model quality checks using MolProbity [ 57 ]. Predictions of EDEM3 posttranslational modification (PTM) were performed with MusiteDeep [ 58 ] and the N-glycans experimentally reported to be attached to the protein core were modelled as described in [ 59 ] using structural data from the SAGS database [ 60 , 61 , 62 ].…”

Section: Methodsmentioning

confidence: 99%

EDEM3 Domains Cooperate to Perform Its Overall Cell Functioning

Manica

Ghenea

Munteanu³

et al. 2021

IJMS

View full text Add to dashboard Cite

EDEM3 recognizes and directs misfolded proteins to the ER-associated protein degradation (ERAD) process. EDEM3 was predicted to act as lectin or as a mannosidase because of its homology with the GH47 catalytic domain of the Man1B1, but the contribution of the other regions remained unresolved. Here, we dissect the molecular determinants governing EDEM3 function and its cellular interactions. LC/MS analysis indicates very few stable ER interactors, suggesting EDEM3 availability for transient substrate interactions. Sequence analysis reveals that EDEM3 consists of four consecutive modules defined as GH47, intermediate (IMD), protease-associated (PA), and intrinsically disordered (IDD) domain. Using an EDEM3 knock-out cell line, we expressed EDEM3 and domain deletion mutants to address EDEM3 function. We find that the mannosidase domain provides substrate binding even in the absence of mannose trimming and requires the IMD domain for folding. The PA and IDD domains deletions do not impair the trimming, but specifically modulate the turnover of two misfolded proteins, NHK and the soluble tyrosinase mutant. Hence, we demonstrate that EDEM3 provides a unique ERAD timing to misfolded glycoproteins, not only by its mannose trimming activity, but also by the positive and negative feedback modulated by the protease-associated and intrinsically disordered domain, respectively.

show abstract

“…[ 117 ] Most recently, MusiteDeep has been extended to incorporate the CapsNet with ensemble techniques for the prediction of more types of PTM sites. [ 128 ] The tool could be easily extended to predict more PTMs given enough number of known sites for training.…”

Section: Deep Learning For Post‐translational Modification Predictionmentioning

confidence: 99%

Deep Learning in Proteomics

et al. 2020

View full text Add to dashboard Cite

Proteomics, the study of all the proteins in biological systems, is becoming a data-rich science. Protein sequences and structures are comprehensively catalogued in online databases. With recent advancements in tandem mass spectrometry (MS) technology, protein expression and post-translational modifications (PTMs) can be studied in a variety of biological systems at the global scale. Sophisticated computational algorithms are needed to translate the vast amount of data into novel biological insights. Deep learning automatically extracts data representations at high levels of abstraction from data, and it thrives in data-rich scientific research domains. Here, a comprehensive overview of deep learning applications in proteomics, including retention time prediction, MS/MS spectrum prediction, de novo peptide sequencing, PTM prediction, major histocompatibility complex-peptide binding prediction, and protein structure prediction, is provided. Limitations and the future directions of deep learning in proteomics are also discussed. This review will provide readers an overview of deep learning and how it can be used to analyze proteomics data.

show abstract

MusiteDeep: a deep-learning based webserver for protein post-translational modification site prediction and visualization

Cited by 171 publications

References 31 publications

Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks

Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks

EDEM3 Domains Cooperate to Perform Its Overall Cell Functioning

Deep Learning in Proteomics

Contact Info

Product

Resources

About