Biochimica et Biophysica Acta 1844 (2014) 163–164
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap
Preface
Computational proteomics, systems biology and clinical implications
With the rapid accumulation of omics data in Proteomics and Genomics, many previous unsolvable problems become easier to study. Accompanied with novel methods that can be applied onto these merging omics data, it may be possible to provide a feasible solution to these complex and difficult biological and clinical problems. In the special issue, many topics were investigated and novel analysis methods were proposed. Chen et al. study the performances of various feature encoding methods based on primary sequence information to qualitatively predict PDZ domain-peptide binding. They then propose a new feature selection algorithm to choose relevant and irredundant sequence-based features to improve PDZ domain-peptide interaction prediction accuracy. Many critical sequence motifs are analyzed and confirmed by prior findings. Liu et al. present a novel structure-based protein zinc-binding site predictor with a geometric restriction method. Large-scale prediction with this powerful predictor indicated that zinc-binding proteins were significantly involved in more complicated biological processes in higher species. Furthermore, it was observed that zinc-binding proteins are preferentially implicated in diseases and highly enriched in potential drug targets, while prediction of potential zinc-binding sites can be helpful for further studies of molecular mechanisms. Kurgan et al. developed a first-of-its-kind high-throughput predictor of cyclic proteins, which are promising targets for pharmaceutical/ therapeutic applications. Their CyPred method accurately finds chains from several cyclic families: cyclotides, cyclic defensins, bacteriocins, and trypsin inhibitors. They used CyPred to find and characterize ~3500 putative cyclic proteins among 5.7+ million chains from 642 fully sequenced proteomes. Ji et al. study the advantages and disadvantages of various methods in microRNA (miRNA) family classification. Particularly, they discuss the impact of miRNA on protein activity and its clinical application in cancer research in a view of miRNA family. The p53 protein plays a central role in the cell by integrating pathways related to apoptosis, cell cycle arrest, and DNA repair, and p53 has the highest cancer mutation prevalence. p53 family proteins (p53, p63 and p73) are structurally and functionally highly similar to p53, particularly in transactivation of similar genes and in maintaining similar interaction network. Ma et al. analyzed the amino acid and dipeptide composition of p53/p63/p73 proteins across evolution and compared them with the gain/loss of amino acids and dipeptides in human p53 following cancer-related somatic mutations. They found that dipeptide mutational gain/loss ratios are inversely correlated with those observed over p53 evolution but tend to follow the increasing p63/p73-like dipeptide propensities. The results revealed that the p53 mutation spectrum is dominated not only by p53 evolution but also by reversal of evolution to a certain degree. 1570-9639/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.bbapap.2013.10.001
Feng et al. proposed a novel computational approach to predict drug target group. The proposed method integrated various information of drug compound including chemical-chemical similarities, chemicalchemical connections, and chemical-protein connections. The results suggest that the target proteins of similar drugs or connected drugs are often in the same target group. Lu et al. applied physicochemical molecular descriptors and pseudo amino acid composition to encode drug and enzyme molecules, respectively. Then, an optimal feature set containing 129 features was selected based on a feature selection method called CfsSubset. Finally, Random Forest was employed to build the drug-enzyme interaction network. The predicting model achieved an MCC of 0.915 and a Sensitivity of 87.9% at the Specificity level of 99.8% for 10-fold cross-validation test, and achieved an MCC of 0.895 and a Sensitivity of 95.7% at the Specificity level of 95.4% for independent set test. It is further found that Geometry features were the most important. Qian et al. performed a comparative analysis of phosphorylation networks between two distant species, human and yeast. Their analysis unveiled a conserved backbone, which is dominant by kinase-to-kinaserelationship. Orthologous kinases share similar functions despite the fact that they target distinct substrate sets. More importantly, their analysis also indicated that, different regulatory mechanisms, such as phosphorylation and transcription, do inherently wire together. Clairambault et al. applied deterministic ordinary differential equation (ODE) method to model ATM/p53/Mdm2/Wip1 intracellular dynamics following DNA damage. Their results demonstrated that p53 oscillations that have been observed in individual cells can be reconstructed and predicted by compartmentalizing cellular events occurring after DNA damage, either in the nucleus or in the cytoplasm. Liu et al. developed a mathematical model to investigate how cell fate decisions are well coordinated by transcription factors and miRNAs. By analyzing nonlinear dynamic behaviors of the model, they found that miR34 plays a critical role in promoting cell cycle arrest by inhibiting E2F. Moreover, miR449 is necessary for apoptosis due to its inhibition to Sirt1 that can repress p53-dependent apoptosis. These results are consistent with existing experimental observations. The construction of a disease model is a challenge that cannot be addressed in a single step, thus raising the question of how to get started. Voit demonstrates with the example of cystic fibrosis that “mesoscopic” models offer a computational infrastructure that facilitates expansions toward increasingly complex disease models. Xie et al. constructed HBV/HCV viral dysfunctional network in Hepatocellular carcinoma (HCC), in hope of investigating viral infection impact on the change of genome expression and protein interaction in the development of HCC. They found that HBx, the main HBV viral protein, directly acted on the gene groups of cell cycle. On the other hand, multiple important HCV viral proteins including CORE, NS3 and NS5A
164
Preface
acted directly or indirectly on THBS1 in TGF-β pathway. Similar results were validated in independent datasets with more samples. To understand the in vivo mechanism of Traditional Chinese medicine (TCM), Li et al. investigated the global biological characterization of the urine of psoriasis patients with Blood Stasis Syndrome and the therapeutic metabolomic mechanism of the Optimized Yinxieling formula with the metabolomic technique combining molecular docking analysis. Metabolomic method for urinary metabolic fingerprinting of Blood Stasis Syndrome of Psoriasis Vulgar was established. The effect of Optimized Yinxieling was investigated using metabonomic study and molecular docking analysis. The method presented in this study might be a complement to further TCM syndrome research. Vera et al. discuss the use of hybrids models, composed of ODE and logical sub-modules, as a strategy to handle large scale, non-linear biochemical networks that include transcriptional and post-transcriptional regulation. Their method is exemplified with a regulatory network centered on E2F1, a transcription factor involved in cancer.
Yudong Cai is a full time professor and principle investigator at the Institute of Systems Biology, Shanghai University, China. He is a member of the Editorial Board of Biochimica et Biophysica Acta: Protein and Proteomics. He is also a guest editor for special issues of Biochimica et Biophysica Acta: Protein and Proteomics on “Computational Proteomics, Systems Biology and Clinical Implications”. His main interests cover various areas of systems biology and bioinformatics such as protein-protein interaction, disease biomarker prediction, drug-target interaction and protein functional site prediction. He has published about 150 papers in peer review journals, with h-index of 42.
Yudong Cai Institute of Systems Biology, Shanghai University, Shanghai 200444, China
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap
Preface
Computational proteomics, systems biology and clinical implications
With the rapid accumulation of omics data in Proteomics and Genomics, many previous unsolvable problems become easier to study. Accompanied with novel methods that can be applied onto these merging omics data, it may be possible to provide a feasible solution to these complex and difficult biological and clinical problems. In the special issue, many topics were investigated and novel analysis methods were proposed. Chen et al. study the performances of various feature encoding methods based on primary sequence information to qualitatively predict PDZ domain-peptide binding. They then propose a new feature selection algorithm to choose relevant and irredundant sequence-based features to improve PDZ domain-peptide interaction prediction accuracy. Many critical sequence motifs are analyzed and confirmed by prior findings. Liu et al. present a novel structure-based protein zinc-binding site predictor with a geometric restriction method. Large-scale prediction with this powerful predictor indicated that zinc-binding proteins were significantly involved in more complicated biological processes in higher species. Furthermore, it was observed that zinc-binding proteins are preferentially implicated in diseases and highly enriched in potential drug targets, while prediction of potential zinc-binding sites can be helpful for further studies of molecular mechanisms. Kurgan et al. developed a first-of-its-kind high-throughput predictor of cyclic proteins, which are promising targets for pharmaceutical/ therapeutic applications. Their CyPred method accurately finds chains from several cyclic families: cyclotides, cyclic defensins, bacteriocins, and trypsin inhibitors. They used CyPred to find and characterize ~3500 putative cyclic proteins among 5.7+ million chains from 642 fully sequenced proteomes. Ji et al. study the advantages and disadvantages of various methods in microRNA (miRNA) family classification. Particularly, they discuss the impact of miRNA on protein activity and its clinical application in cancer research in a view of miRNA family. The p53 protein plays a central role in the cell by integrating pathways related to apoptosis, cell cycle arrest, and DNA repair, and p53 has the highest cancer mutation prevalence. p53 family proteins (p53, p63 and p73) are structurally and functionally highly similar to p53, particularly in transactivation of similar genes and in maintaining similar interaction network. Ma et al. analyzed the amino acid and dipeptide composition of p53/p63/p73 proteins across evolution and compared them with the gain/loss of amino acids and dipeptides in human p53 following cancer-related somatic mutations. They found that dipeptide mutational gain/loss ratios are inversely correlated with those observed over p53 evolution but tend to follow the increasing p63/p73-like dipeptide propensities. The results revealed that the p53 mutation spectrum is dominated not only by p53 evolution but also by reversal of evolution to a certain degree. 1570-9639/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.bbapap.2013.10.001
Feng et al. proposed a novel computational approach to predict drug target group. The proposed method integrated various information of drug compound including chemical-chemical similarities, chemicalchemical connections, and chemical-protein connections. The results suggest that the target proteins of similar drugs or connected drugs are often in the same target group. Lu et al. applied physicochemical molecular descriptors and pseudo amino acid composition to encode drug and enzyme molecules, respectively. Then, an optimal feature set containing 129 features was selected based on a feature selection method called CfsSubset. Finally, Random Forest was employed to build the drug-enzyme interaction network. The predicting model achieved an MCC of 0.915 and a Sensitivity of 87.9% at the Specificity level of 99.8% for 10-fold cross-validation test, and achieved an MCC of 0.895 and a Sensitivity of 95.7% at the Specificity level of 95.4% for independent set test. It is further found that Geometry features were the most important. Qian et al. performed a comparative analysis of phosphorylation networks between two distant species, human and yeast. Their analysis unveiled a conserved backbone, which is dominant by kinase-to-kinaserelationship. Orthologous kinases share similar functions despite the fact that they target distinct substrate sets. More importantly, their analysis also indicated that, different regulatory mechanisms, such as phosphorylation and transcription, do inherently wire together. Clairambault et al. applied deterministic ordinary differential equation (ODE) method to model ATM/p53/Mdm2/Wip1 intracellular dynamics following DNA damage. Their results demonstrated that p53 oscillations that have been observed in individual cells can be reconstructed and predicted by compartmentalizing cellular events occurring after DNA damage, either in the nucleus or in the cytoplasm. Liu et al. developed a mathematical model to investigate how cell fate decisions are well coordinated by transcription factors and miRNAs. By analyzing nonlinear dynamic behaviors of the model, they found that miR34 plays a critical role in promoting cell cycle arrest by inhibiting E2F. Moreover, miR449 is necessary for apoptosis due to its inhibition to Sirt1 that can repress p53-dependent apoptosis. These results are consistent with existing experimental observations. The construction of a disease model is a challenge that cannot be addressed in a single step, thus raising the question of how to get started. Voit demonstrates with the example of cystic fibrosis that “mesoscopic” models offer a computational infrastructure that facilitates expansions toward increasingly complex disease models. Xie et al. constructed HBV/HCV viral dysfunctional network in Hepatocellular carcinoma (HCC), in hope of investigating viral infection impact on the change of genome expression and protein interaction in the development of HCC. They found that HBx, the main HBV viral protein, directly acted on the gene groups of cell cycle. On the other hand, multiple important HCV viral proteins including CORE, NS3 and NS5A
164
Preface
acted directly or indirectly on THBS1 in TGF-β pathway. Similar results were validated in independent datasets with more samples. To understand the in vivo mechanism of Traditional Chinese medicine (TCM), Li et al. investigated the global biological characterization of the urine of psoriasis patients with Blood Stasis Syndrome and the therapeutic metabolomic mechanism of the Optimized Yinxieling formula with the metabolomic technique combining molecular docking analysis. Metabolomic method for urinary metabolic fingerprinting of Blood Stasis Syndrome of Psoriasis Vulgar was established. The effect of Optimized Yinxieling was investigated using metabonomic study and molecular docking analysis. The method presented in this study might be a complement to further TCM syndrome research. Vera et al. discuss the use of hybrids models, composed of ODE and logical sub-modules, as a strategy to handle large scale, non-linear biochemical networks that include transcriptional and post-transcriptional regulation. Their method is exemplified with a regulatory network centered on E2F1, a transcription factor involved in cancer.
Yudong Cai is a full time professor and principle investigator at the Institute of Systems Biology, Shanghai University, China. He is a member of the Editorial Board of Biochimica et Biophysica Acta: Protein and Proteomics. He is also a guest editor for special issues of Biochimica et Biophysica Acta: Protein and Proteomics on “Computational Proteomics, Systems Biology and Clinical Implications”. His main interests cover various areas of systems biology and bioinformatics such as protein-protein interaction, disease biomarker prediction, drug-target interaction and protein functional site prediction. He has published about 150 papers in peer review journals, with h-index of 42.
Yudong Cai Institute of Systems Biology, Shanghai University, Shanghai 200444, China
