Improved Prediction of Cell-Penetrating Peptides via Effective Orchestrating Amino Acid Composition Feature Representation

Fu, Xiquan; Ke, Lixia; Cai, Liang; Chen, Xiangtao; Ren, Xin; Gao, Mingyu

doi:10.1109/access.2019.2952738

Cited by 16 publications

(10 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to distinguish between thermophilic and nonthermophilic proteins, SVM (Ding et al, 2016a,b;He et al, 2018;Qiao et al, 2018;Wei et al, 2018;Fu et al, 2019;Wang et al, 2019b), random forest [RF, (Ding et al, 2017;Wang et al, 2019a)], decision tree (Mohasseb et al, 2018;, and naïve Bayes [NB, (Rajaraman and Chokkalingam, 2014)] methods were used in our experiment. The first two methods were implemented and optimized in the python 3.7 environment with our edited code.…”

Section: Machine Learning Methodsmentioning

confidence: 99%

“…To quantitatively identify proteins, the physicochemical characteristics were obtained using a method (temporarily called 188d), which could extract sequence information and amino acid properties (Song et al, 2014;Xu et al, 2014Xu et al, , 2018Fu et al, 2019;Liu, 2019;Zhu et al, 2019). The first 20 elements in the results of this method denoted the frequency of the 20 original amino acids (Zhu et al, 2019); the next 24 features reflected the group proportion corresponding to three groups (Qu et al, 2019); the following 120 dimensions were the distributions of three groups in five local positions (Cai et al, 2003); the last 24 features were the numbers of three types of dipeptides.…”

Section: Physicochemical Characteristicsmentioning

confidence: 99%

See 1 more Smart Citation

A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

Feng

Dan

et al. 2020

Front. Bioeng. Biotechnol.

View full text Add to dashboard Cite

The thermostability of proteins is a key factor considered during enzyme engineering, and finding a method that can identify thermophilic and non-thermophilic proteins will be helpful for enzyme design. In this study, we established a novel method combining mixed features and machine learning to achieve this recognition task. In this method, an amino acid reduction scheme was adopted to recode the amino acid sequence. Then, the physicochemical characteristics, auto-cross covariance (ACC), and reduced dipeptides were calculated and integrated to form a mixed feature set, which was processed using correlation analysis, feature selection, and principal component analysis (PCA) to remove redundant information. Finally, four machine learning methods and a dataset containing 500 random observations out of 915 thermophilic proteins and 500 random samples out of 793 non-thermophilic proteins were used to train and predict the data. The experimental results showed that 98.2% of thermophilic and non-thermophilic proteins were correctly identified using 10-fold cross-validation. Moreover, our analysis of the final reserved features and removed features yielded information about the crucial, unimportant and insensitive elements, it also provided essential information for enzyme design.

show abstract

Section: Machine Learning Methodsmentioning

confidence: 99%

Section: Physicochemical Characteristicsmentioning

confidence: 99%

A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

Feng

Dan

et al. 2020

Front. Bioeng. Biotechnol.

View full text Add to dashboard Cite

show abstract

“…In our experiment, we used the following four indicators to evaluate the predictive performance of our proposed model, including Accuracy (ACC), Sensitivity (SN), Specificity (SP), and Mathew's Correlation Coefficient (MCC). They are the four commonly used indicators for classifier performance evaluation in other Bioinformatics fields (Zhang et al, 2008(Zhang et al, , 2018a(Zhang et al, ,b,c, 2019bWei et al, 2017bWei et al, , 2019bZeng et al, 2017bZeng et al, , 2019cChen et al, 2018;Lu et al, 2018a,b;Fu et al, 2019;Gong et al, 2019;Jin et al, 2019;Liu and Li, 2019;Liu et al, 2019c,d;Manavalan et al, 2019a,b,c,d;Basith et al, 2020). Their calculation formulas are as follows:…”

Section: Performance Indicatorsmentioning

confidence: 99%

Developing a Multi-Layer Deep Learning Based Predictive Model to Identify DNA N4-Methylcytosine Modifications

Zeng

Liao

2020

Front. Bioeng. Biotechnol.

View full text Add to dashboard Cite

DNA N4-methylcytosine modification (4mC) plays an essential role in a variety of biological processes. Therefore, accurate identification the 4mC distribution in genome-scale is important for systematically understanding its biological functions. In this study, we present Deep4mcPred, a multi-layer deep learning based predictive model to identify DNA N4-methylcytosine modifications. In this predictor, we for the first time integrate residual network and recurrent neural network to build a multi-layer deep learning predictive system. As compared to existing predictors using traditional machine learning, our proposed method has two advantages. First, our deep learning framework does not need to specify the features when training the predictive model. It can automatically learn the high-level features and capture the characteristic specificity of 4mC sites, benefiting to distinguish true 4mC sites from non-4mC sites. On the other hand, our deep learning method outperforms the traditional machine learning predictors in performance by benchmarking comparison, demonstrating that the proposed Deep4mcPred is more effective in the DNA 4mC site prediction. Moreover, via experimental comparison, we found that attention mechanism introduced into the deep learning framework is useful to capture the critical features. Additionally, we develop a webserver implementing the proposed method for the academic use of research community, which is now available at http://server.malab.cn/Deep4mcPred.

show abstract

“…Since safety, stability, and specificity of peptides for various utilizations is of great value, there is always the need to introduce novel natural or synthetic CPPs. As protein derived penetrating peptides constitute about 80–90% of CPPs, it is arguable that huge quantities of CPPs are hidden in protein sequences ( Fu et al, 2019 ). Screening the proteome of the Coronaviridae family for this purpose was somehow neglected until the recent pandemic.…”

Section: Introductionmentioning

confidence: 99%

“…There are various indicators such as sensitivity, specificity, accuracy, receiver operating characteristic, and Matthew's correlation coefficient to assess the performance of the models. Then statistical methods such as jackknife test, k-fold cross validation, and independent test are used to verify indicators in different predictor models ( Fu et al, 2019 ).…”

Section: Introductionmentioning

confidence: 99%

Decoding the proteome of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) for cell-penetrating peptides involved in pathogenesis or applicable as drug delivery vectors

Hemmati

Behzadipour

Haddad

2020

Infection, Genetics and Evolution

View full text Add to dashboard Cite

Synthetic or natural derived cell-penetrating peptides (CPPs) are vastly investigated as tools for the intracellular delivery of membrane-impermeable molecules. As viruses are intracellular obligate parasites, viral originated CPPs have been considered as suitable intracellular shuttling vectors for cargo transportation. A total of 310 CPPs were identified in the proteome of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Screening the proteome of the cause of COVID-19 reveals that SARS-CoV-2 CPPs (SCV2-CPPs) span the regions involved in replication, protein-nucleotide and protein-protein interaction, protein-metal ion interaction, and stabilization of homo/hetero-oligomers. However, to find the most appropriate peptides as drug delivery vectors, one might face several hurdles. Computational analyses showed that 94.3% of the identified SCV2-CPPs are non-toxins, and 38% are neither antigenic nor allergenic. Interestingly, 36.70% of SCV2-CPPs were resistant to all four groups of protease families. Nearly 1/3 of SCV2-CPPs had sufficient inherent or induced helix and sheet conformation leading to increased uptake efficiency. Heliquest lipid-binding discrimination factor revealed that 44.30% of the helical SCV2-CPPs are lipid-binding helices. Although Cys-rich derived CPPs of helicase (NSP13) can potentially fold into a cyclic conformation in endosomes with a higher rate of endosomal release, the most optimal SCV2-CPP candidates as vectors for drug delivery were SCV2-CPP118, SCV2-CPP119, SCV2-CPP122, and SCV2-CPP129 of NSP12 (RdRp). Ten experimentally validated viral-derived CPPs were also used as the positive control to check the scalability and reliability of our protocol in SCV2-CPP retrieval. Some peptides with a cell-penetration ability known as bioactive peptides are adopted as biotherapeutics themselves. Therefore, 59.60%, 29.63%, and 32.32% of SCV2-CPPs were identified as potential antibacterial, antiviral, and antifungals, respectively. While 63.64% of SCV2-CPPs had immuno-modulatory properties, 21.89% were recognized as anti-cancers. Conclusively, the workflow of this study provides a platform for profound screening of viral proteomes as a rich source of biotherapeutics or drug delivery carriers.

show abstract

Improved Prediction of Cell-Penetrating Peptides via Effective Orchestrating Amino Acid Composition Feature Representation

Cited by 16 publications

References 97 publications

A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

A Method for Prediction of Thermophilic Protein Based on Reduced Amino Acids and Mixed Features

Developing a Multi-Layer Deep Learning Based Predictive Model to Identify DNA N4-Methylcytosine Modifications

Decoding the proteome of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) for cell-penetrating peptides involved in pathogenesis or applicable as drug delivery vectors

Contact Info

Product

Resources

About