Limits...
CleavPredict: A Platform for Reasoning about Matrix Metalloproteinases Proteolytic Events.

Kumar S, Ratnikov BI, Kazanov MD, Smith JW, Cieplak P - PLoS ONE (2015)

Bottom Line: The prediction method employs the MMP specific position weight matrices (PWMs) derived from statistical analysis of high-throughput phage display experimental results.All this information will provide the user with perspectives in reasoning about proteolytic events.CleavPredict is freely accessible, and there is no login required.

View Article: PubMed Central - PubMed

Affiliation: Sanford Burnham Medical Research Institute, La Jolla, California, United States of America.

ABSTRACT
CleavPredict (http://cleavpredict.sanfordburnham.org) is a Web server for substrate cleavage prediction for matrix metalloproteinases (MMPs). It is intended as a computational platform aiding the scientific community in reasoning about proteolytic events. CleavPredict offers in silico prediction of cleavage sites specific for 11 human MMPs. The prediction method employs the MMP specific position weight matrices (PWMs) derived from statistical analysis of high-throughput phage display experimental results. To augment the substrate cleavage prediction process, CleavPredict provides information about the structural features of potential cleavage sites that influence proteolysis. These include: secondary structure, disordered regions, transmembrane domains, and solvent accessibility. The server also provides information about subcellular location, co-localization, and co-expression of proteinase and potential substrates, along with experimentally determined positions of single nucleotide polymorphism (SNP), and posttranslational modification (PTM) sites in substrates. All this information will provide the user with perspectives in reasoning about proteolytic events. CleavPredict is freely accessible, and there is no login required.

No MeSH data available.


ROC curves for prediction cleavage sites in proteins collected in CutDB for MMP-2, MMP-3, MMP-8, MMP-9 and MMP-14.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4440711&req=5

pone.0127877.g003: ROC curves for prediction cleavage sites in proteins collected in CutDB for MMP-2, MMP-3, MMP-8, MMP-9 and MMP-14.

Mentions: Further on we validated our PWM—based scoring method in another “external” test performed on substrates collected in CutDB [3]. We selected only those MMPs for which sufficient number of cleavage events and protein substrates is available. Thus, we applied our prediction algorithm to calculate scores for cleavage sites in substrates of five MMPs including: MMP-9 (334 cleavages in 88 unique substrates), MMP-2 (135 cleavages in 50 substrates), MMP-14 (89 cleavages in 38 substrates), MMP-3 (186 cleavages in 67 proteins) and MMP-8 (85 cleavages in 26 proteins). In these calculations, the positive set constitutes the cleavage sites reported in the literature (CutDB) for appropriate protein substrates for each MMP, while for the negative set we choose peptide bonds randomly selected from the same protein substrates that are not cleaved by MMP. The ratio of positive to negative cases is 1:100. The results of our prediction calculations are collected in Table 4. It demonstrates that for experimental protein substrates the PWM approach yields the accuracy reaching the level of 70% for most MMPs, while the false-positive rate is in the range of 30%, with the exception of MMP-8 and MMP-9 for which false-positive rate is 61 and 47%, respectively. The appropriate ROC curves are presented in Fig 3. Area under the curve (AUC) (Table 4), in most cases is well above 0.8, which demonstrate a good ability of our method to discriminate between cleavable and non-cleavable peptide bonds for MMP hydrolysis. The high level of false-positive rate is not satisfactory here and substantially higher than for uniformly identified substrates by Overall et al., as discussed above. However, we are aware that the reported cleavage sites come from highly heterogeneous sources and may not all be entirely accurate, either because denatured proteins were used as substrates, or because when the study was performed the methods for determining the position of the cleavage sites, including mass spectroscopy, were not as robust as methods available today. What is more important, the conditions used for studying cleavage events reported in the literature could differ substantially from those used in our phage display experiment. The conditions used in our high throughput phage display experiment allow measuring important cleavage events with observed kcat/KM values above 3000 sec-1M-1 [53]. Thus, if our predictive algorithm is sufficiently accurate, we may be able to identify the reported cleavage sites that are “suspect.”


CleavPredict: A Platform for Reasoning about Matrix Metalloproteinases Proteolytic Events.

Kumar S, Ratnikov BI, Kazanov MD, Smith JW, Cieplak P - PLoS ONE (2015)

ROC curves for prediction cleavage sites in proteins collected in CutDB for MMP-2, MMP-3, MMP-8, MMP-9 and MMP-14.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4440711&req=5

pone.0127877.g003: ROC curves for prediction cleavage sites in proteins collected in CutDB for MMP-2, MMP-3, MMP-8, MMP-9 and MMP-14.
Mentions: Further on we validated our PWM—based scoring method in another “external” test performed on substrates collected in CutDB [3]. We selected only those MMPs for which sufficient number of cleavage events and protein substrates is available. Thus, we applied our prediction algorithm to calculate scores for cleavage sites in substrates of five MMPs including: MMP-9 (334 cleavages in 88 unique substrates), MMP-2 (135 cleavages in 50 substrates), MMP-14 (89 cleavages in 38 substrates), MMP-3 (186 cleavages in 67 proteins) and MMP-8 (85 cleavages in 26 proteins). In these calculations, the positive set constitutes the cleavage sites reported in the literature (CutDB) for appropriate protein substrates for each MMP, while for the negative set we choose peptide bonds randomly selected from the same protein substrates that are not cleaved by MMP. The ratio of positive to negative cases is 1:100. The results of our prediction calculations are collected in Table 4. It demonstrates that for experimental protein substrates the PWM approach yields the accuracy reaching the level of 70% for most MMPs, while the false-positive rate is in the range of 30%, with the exception of MMP-8 and MMP-9 for which false-positive rate is 61 and 47%, respectively. The appropriate ROC curves are presented in Fig 3. Area under the curve (AUC) (Table 4), in most cases is well above 0.8, which demonstrate a good ability of our method to discriminate between cleavable and non-cleavable peptide bonds for MMP hydrolysis. The high level of false-positive rate is not satisfactory here and substantially higher than for uniformly identified substrates by Overall et al., as discussed above. However, we are aware that the reported cleavage sites come from highly heterogeneous sources and may not all be entirely accurate, either because denatured proteins were used as substrates, or because when the study was performed the methods for determining the position of the cleavage sites, including mass spectroscopy, were not as robust as methods available today. What is more important, the conditions used for studying cleavage events reported in the literature could differ substantially from those used in our phage display experiment. The conditions used in our high throughput phage display experiment allow measuring important cleavage events with observed kcat/KM values above 3000 sec-1M-1 [53]. Thus, if our predictive algorithm is sufficiently accurate, we may be able to identify the reported cleavage sites that are “suspect.”

Bottom Line: The prediction method employs the MMP specific position weight matrices (PWMs) derived from statistical analysis of high-throughput phage display experimental results.All this information will provide the user with perspectives in reasoning about proteolytic events.CleavPredict is freely accessible, and there is no login required.

View Article: PubMed Central - PubMed

Affiliation: Sanford Burnham Medical Research Institute, La Jolla, California, United States of America.

ABSTRACT
CleavPredict (http://cleavpredict.sanfordburnham.org) is a Web server for substrate cleavage prediction for matrix metalloproteinases (MMPs). It is intended as a computational platform aiding the scientific community in reasoning about proteolytic events. CleavPredict offers in silico prediction of cleavage sites specific for 11 human MMPs. The prediction method employs the MMP specific position weight matrices (PWMs) derived from statistical analysis of high-throughput phage display experimental results. To augment the substrate cleavage prediction process, CleavPredict provides information about the structural features of potential cleavage sites that influence proteolysis. These include: secondary structure, disordered regions, transmembrane domains, and solvent accessibility. The server also provides information about subcellular location, co-localization, and co-expression of proteinase and potential substrates, along with experimentally determined positions of single nucleotide polymorphism (SNP), and posttranslational modification (PTM) sites in substrates. All this information will provide the user with perspectives in reasoning about proteolytic events. CleavPredict is freely accessible, and there is no login required.

No MeSH data available.