In this study, we proposed a hybrid deep-learning algorithm named DeepOMe that combined Convolutional Neural Networks (CNN) and Bidirectional Long Short-term Memory (BLSTM) to accurately predict 2′-O-Me sites in human transcriptome. Validating under 4-, 6-, 8-, and 10-fold cross-validation, we confirmed that our proposed model achieved a high performance (AUC close to 0.998 and AUPR close to 0.880). To facilitate the usage of DeepOMe, a user-friendly web-server was constructed, which can be freely accessed at http://deepome.renlab.org.

autoRPA is an online service that can construct prognostic staging models and perform performance comparison between different staging models. By using Recursive Partitioning Analysis (RPA) and the log-rank test statistics, autoRPA will provide an intuitive way for building decision-making tree from survival data. The web server of autoRPA is implemented in JAVA and PHP and is freely available at: http://rpa.renlab.org/.

In this study, we develop an online service that annotates PTM alterations from genetic mutations data such as germline SNPs or cancer somatic mutations. We present the analysis result in several interactive charts and conducted GO and KEGG enrichment analysis on the identified significantly PTM-mutated proteins. The web server of PTMsnp is implemented in Python and PHP and is freely available at: http://ptmsnp.renlab.org/.

In this study, we aimed to develop a computational tool for predicting nitration and nitrosylation sites in proteins. We first constructed four novel types of encoding features to represent the modified residues. Then, a predictor called DeepNitro was established using deep learning methods. The web server of DeepNitro is implemented in JAVA and PHP and is freely available at: http://deepnitro.renlab.org.

In this study, we present a user-friendly web server called “m6ASNP” that is dedicated to the identification of genetic variants targeting m6A modification sites.In m6ASNP, genetic variants in a standard VCF format are accepted as the input data, and the output includes an interactive table containing the genetic variants annotated by m6A function. The web server “m6ASNP” is implemented in JAVA and PHP and is freely available at: http://m6asnp.renlab.org

Computational prediction of phosphorylation sites with their cognate protein kinases (PKs) is greatly helpful for further experimental design. Although ~10 online predictors were developed, the PK classification and control of false positive rate (FPR) were not well addressed. Here we adopted a well-established rule to classify PKs into a hierarchical structure with four levels. GPS 2.0 is freely available at: http://sumosp.biocuckoo.org

As a special class of post-translational modifications (PTMs), numerous proteins could be covalently modified by a variety of lipids, including myristate (C14), palmitate (C16), farnesyl (C15), geranylgeranyl (C20) and glycosylphosphatidylinositol (GPI), etc.During the past decade, we have developed a series of high-performance protein palmitoylation sites predictors. The online service and local packages of CSS-Palm 2.0 were freely available at: http://bioinformatics.lcd-ustc.org/css_palm

Recently, it was reported SUMOs can non-covalently interact with other proteins through targeting specific SIMs.In this work, by improving the prediction algorithm and adding the novel SIMs prediction feature, we developed an updated version of SUMOsp and renamed it as GPS-SUMO. From the scientific literature, we manually collected 983 sumoylation sites in 545 proteins and 151 known SIMs in 80 proteins as the non-redundant data sets, respectively. The online service and local packages of SUMOsp 2.0 are freely available at: http://sumosp.biocuckoo.org

The irregular lipid modification with essential physiological functions, may lead to all sorts of diseases.In this work, we present GPS-Lipid, which is a comprehensive predictor for protein lipid modification sites. From the literatures published before November, 2014, we manually collected 737 S-palmitoylation sites in 361 proteins, 106 S-farnesylation sites in 97 proteins, 95 S- geranylgeranylation sites in 70 proteins and 283 N-myristoylation sites in 281 proteins. we hosted a publicly available web server at: http://lipid.biocuckoo.org

Recently, prokaryotic ubiquitin-like protein (PUP) was identified as the tagging system in prokaryotes (Pearce, M. J. et al., 2008).In this work, we manually collected 127 experimentally indentified protein pupylation sites in 109 unique proteins from scientific literature. We calculated the leave-one-out validation and 4-, 6-, 8-, 10-fold cross-validations to evaluate the prediction performance and system robustness. The leave-one-out validation result is accuracy (Ac) of 78.85%, sensitivity (Sn) of 63.78%, and specificity (Sp) of 80.21%. The GPS-PUP 1.0 is freely available at: http://pup.biocuckoo.org

In this work, we manually collected 467 experimentally verified S-nitrosylation sites in 302 unique proteins from scientific literature. Previously, we developed an algorithm of GPS 2.0 (Group-based Prediction System) for prediction of kinase-specific phosphorylation sites (Xue, et al., 2008). Here, we greatly improved the method and released GPS 3.0 algorithm. Then we developed a novel computational software of GPS-SNO 1.0 for prediction of S-nitrosylation sites. The online service and local packages of GPS-SNO were implemented in JAVA and are freely available at: http://sno.biocuckoo.org/

In this work, we manually collected 1066 experimentally indentified protein nitration sites in 554 unique proteins from scientific literature. A previously self-developed GPS (Group-based Prediction System) algorithm was employed with great improvement. We calculated the leave-one-out validation and 4-, 6-, 8-, 10-fold cross-validations to evaluate the prediction performance and system robustness. The leave-one-out validation result is accuracy (Ac) of 76.51%, sensitivity (Sn) of 50.09%, and specificity (Sp) of 80.18%. The online service and local packages of GPS-YNO2 1.0 were implemented in JAVA and freely available at: http://yno2.biocuckoo.org/

In this work, we collected 368 experimentally verified calpain cleavage sites in 130 proteins. With a previously developed algorithm of GPS (Group-based Prediction System)(Xue et al, 2008), we developed a novel software package of GPS-CCD (Calpain Cleavage Detector) for the prediction of calpain cleavage sites. As an application, we predicted potential calpain cleavage sites for these targets. These prediction results might be a useful resource for further experimental investigation. The online service and local packages of GPS-CCD 1.0 were implemented in JAVA and are freely available at: http://ccd.biocuckoo.org/

In this work, we manually collected experimentally verified 56 phospho-binding sites in 47 distinct substrates and 275 phosphoylation sites in 124 unique proteins from scientific literatures. We calculated the leave-one-out validation and 4-, 6-, 8-, 10-fold cross-validations to evaluate the prediction performance and system robustness. The leave-one-out validation result for phospho-binding prediction is accuracy (Ac) of 95.27%, sensitivity (Sn) of 82.14%, and specificity (Sp) of 95.41%, while for phosphoylation prediction is accuracy (Ac) of 84.55%, sensitivity (Sn) of 53.62%, and specificity (Sp) of 85.03%. The online service and stand-alone packages of GPS-Polo 1.0 were implemented in JAVA 1.4.2 and freely available at:http://polo.biocuckoo.org/

Tyrosine sulfation is a ubiquitous PTM that predominantly modifies proteins in the secretory pathway , and plays an important role in regulating chemotaxis, inflammatory response, and cell adhesion .In this work, we manually collected 273 experimentally indentified protein tyrosine sulfation sites in 171 unique proteins from scientific literature. We calculated the leave-one-out validation and 4-, 6-, 8-, 10-fold cross-validations to evaluate the prediction performance and system robustness. The leave-one-out validation result is accuracy (Ac) of 90.23%, sensitivity (Sn) of 89.60%, and specificity (Sp) of 90.36%. The online service and stand-alone packages of GPS-TSP 1.0 were implemented in JAVA 1.4.2 and freely available at: http://tsp.biocuckoo.org/

In this work, we manually collected all experimentally indentified D-box or KEN-box proteins from scientific literature. A previously self-developed GPS (Group-based Prediction System) algorithm was employed with great improvement. We calculated the leave-one-out validation and 4-, 6-, 8-, 10-fold cross-validations to evaluate the prediction performance and system robustness. The leave-one-out validation result for KEN boxes recognized by Cdh1 is accuracy (Ac) of 95.00%, sensitivity (Sn) of 93.18%, and specificity (Sp) of 100.00%, while for D-boxes recognized by Cdc20 is 87.29% (Ac), 63.51% (Sn), and 95.39% (Sp). The online service and stand-alone packages of GPS-ARM 1.0 were implemented in JAVA 1.6 and freely available at: http://arm.biocuckoo.org/

In this work, we developed a novel GPS-MBA software package for the prediction of I-Ag7 and HLA-DQ8. The experimentally identified epitopes were collected from the scientific literature, while a modified Gibbs sampling approach was adopted to determine core nonamers in these epitopes. For training and prediction, a refined GPS algorithm was used.All experimentally identified T1D antigens together with their epitopes were taken into the TEDB 1.0. Taken together, the prediction and analysis results are helpful for further experimental consideration, and the GPS-MBA can serve as a useful program for experimentalists. The online service and local packages of GPS-MBA 1.0 were implemented in JAVA and could be freely accessed for academic research at: http://mba.biocuckoo.org/

Frequently, experimental researchers prefer to study novel PTM sites to elucidate new functions. Thus, it's important to show whether there have been any sites experimentally verified as real PTM sites in given proteins.In this work, we developed a novel software of PTMs Peptide Scanner (PPS), to reveal known or highly potential PTM sites in eukaryotic proteins. Five typical PTMs were considered, including phosphorylation, sumoylation, palmitoylation, methylation and acetylation. The PPS 1.0 is freely available at: http://pps.biocuckoo.org/

Protein methylation is one of the most important reversible post-translational modifications (PTMs). In this work, we adopted GPS 3.0 algorithm and built GPS-MSP (Methyl-group Specific Predictor) for the prediction of general or type-specific methylline and methylarginine residues in proteins. We also construced up to 28 organism-specific predictors. The GPS-MSP is freely available at: http://msp.biocuckoo.org/

The illustration of biological sequences has become one of the most essential skills for biologists, who at present mainly use the Microsoft PowerPoint, Adobe Illustrator or Photoshop for diagramming biological sequences.In this work, we present a new tool called Illustrator for Biological Sequences (IBS) for assisting experimentalists in drawing publication-quality diagrams of both protein and nucleotide sequences. Both the standalone package and online service are freely available at http://ibs.biocuckoo.org.

A heatmap is a graphical representation of data where the individual values contained in a matrix are represented as colors. In this work, we present a novel software of HemI (Heatmap Illustrator, version 1.0) for experimentalists, to prepare publication-quality figures of heatmap.The stand-alone packages of HemI were implemented in Java and can be accessed at :http://hemi.biocuckoo.org/down.php

In this work, the sRNA-seq technology was first used to determine the expression profiles of sRNAs during eight stages of early zebrafish embryonic development. Based on known zebrafish pre-miRNAs, we designed a zebrafish-specific algorithm of ZmirP (zebrafish miRNA prediction), with 8 new and 57 previously reported sequence and structure features. we greatly improved the CPSS (Zhang Y, et al., 2012) and developed a more specific platform as CSZ (characterization of small RNAome for zebrafish) for the analysis of the high-through sequencing data.The CSZ package was implemented in Perl and freely downloadable at: http://csz.biocuckoo.org/download/csz_1.0.zip