Limits...
SELPHI: correlation-based identification of kinase-associated networks from global phospho-proteomics data sets.

Petsalaki E, Helbig AO, Gopal A, Pasculescu A, Roth FP, Pawson T - Nucleic Acids Res. (2015)

Bottom Line: While phospho-proteomics studies have shed light on the dynamics of cellular signaling, they mainly describe global effects and rarely explore mechanistic details, such as kinase/substrate relationships.In this data set, SELPHI revealed information overlooked by the reporting study, including the known role of MET and EPHA2 kinases in conferring resistance to erlotinib in TKI sensitive strains.SELPHI can significantly enhance the analysis of phospho-proteomics data contributing to improved understanding of sample-specific signaling networks.

View Article: PubMed Central - PubMed

Affiliation: Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, Ontario, M5G 1X8, Canada petsalakis@lunenfeld.ca.

Show MeSH

Related in: MedlinePlus

Workflow of SELPHI. SELPHI first identifies the UniprotKB IDs and sequence location of the input phosphosites. The data are then filtered according to the input cutoff and clustering, KEGG pathway/GO term enrichment analysis and correlation analysis is applied on the input data. The result is an exploratory representation of the global effects relating to cell pathways and functions, a network view of potentially relevant kinase/phosphatase/substrate associations with the likely flow of signaling, as well as motifs enriched in the data set.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4489257&req=5

Figure 1: Workflow of SELPHI. SELPHI first identifies the UniprotKB IDs and sequence location of the input phosphosites. The data are then filtered according to the input cutoff and clustering, KEGG pathway/GO term enrichment analysis and correlation analysis is applied on the input data. The result is an exploratory representation of the global effects relating to cell pathways and functions, a network view of potentially relevant kinase/phosphatase/substrate associations with the likely flow of signaling, as well as motifs enriched in the data set.

Mentions: SELPHI's workflow is presented in Figure 1. If the user has provided protein sequences, SELPHI first pre-processes the input data to identify the associated UniprotKB ID (17) and phosphosite positions in the sequence and filters the data according to the fold change cutoff defined (default is 3-fold). Duplicate peptide identifications are merged using either their weighted average (if there is an intensity or score column available) or the value that represents the maximum change, according to the preferred user settings. Combined input data sets with less than two data points, defined as a ratio or intensity value acquired by a single sample, replicate and condition, are not processed by SELPHI, while the user can restrict their data to only peptides that appear in a minimum number of samples/conditions through the submission interface. There is no limit to the number of peptides in a data set. KEGG pathway (18) and GO Term (FuncAssociate 5) enrichment for the changed peptides is then calculated. The results are plotted as a clustered matrix of the over-represented pathways/GO terms against the samples (Figure 2A, B), colored by the additive (pathways) or average additive (GO terms) log-ratio of intensity for peptides belonging to that term for each sample and demonstrating the enrichment's effect size (log odds ratio) via the size of the box.


SELPHI: correlation-based identification of kinase-associated networks from global phospho-proteomics data sets.

Petsalaki E, Helbig AO, Gopal A, Pasculescu A, Roth FP, Pawson T - Nucleic Acids Res. (2015)

Workflow of SELPHI. SELPHI first identifies the UniprotKB IDs and sequence location of the input phosphosites. The data are then filtered according to the input cutoff and clustering, KEGG pathway/GO term enrichment analysis and correlation analysis is applied on the input data. The result is an exploratory representation of the global effects relating to cell pathways and functions, a network view of potentially relevant kinase/phosphatase/substrate associations with the likely flow of signaling, as well as motifs enriched in the data set.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4489257&req=5

Figure 1: Workflow of SELPHI. SELPHI first identifies the UniprotKB IDs and sequence location of the input phosphosites. The data are then filtered according to the input cutoff and clustering, KEGG pathway/GO term enrichment analysis and correlation analysis is applied on the input data. The result is an exploratory representation of the global effects relating to cell pathways and functions, a network view of potentially relevant kinase/phosphatase/substrate associations with the likely flow of signaling, as well as motifs enriched in the data set.
Mentions: SELPHI's workflow is presented in Figure 1. If the user has provided protein sequences, SELPHI first pre-processes the input data to identify the associated UniprotKB ID (17) and phosphosite positions in the sequence and filters the data according to the fold change cutoff defined (default is 3-fold). Duplicate peptide identifications are merged using either their weighted average (if there is an intensity or score column available) or the value that represents the maximum change, according to the preferred user settings. Combined input data sets with less than two data points, defined as a ratio or intensity value acquired by a single sample, replicate and condition, are not processed by SELPHI, while the user can restrict their data to only peptides that appear in a minimum number of samples/conditions through the submission interface. There is no limit to the number of peptides in a data set. KEGG pathway (18) and GO Term (FuncAssociate 5) enrichment for the changed peptides is then calculated. The results are plotted as a clustered matrix of the over-represented pathways/GO terms against the samples (Figure 2A, B), colored by the additive (pathways) or average additive (GO terms) log-ratio of intensity for peptides belonging to that term for each sample and demonstrating the enrichment's effect size (log odds ratio) via the size of the box.

Bottom Line: While phospho-proteomics studies have shed light on the dynamics of cellular signaling, they mainly describe global effects and rarely explore mechanistic details, such as kinase/substrate relationships.In this data set, SELPHI revealed information overlooked by the reporting study, including the known role of MET and EPHA2 kinases in conferring resistance to erlotinib in TKI sensitive strains.SELPHI can significantly enhance the analysis of phospho-proteomics data contributing to improved understanding of sample-specific signaling networks.

View Article: PubMed Central - PubMed

Affiliation: Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, Ontario, M5G 1X8, Canada petsalakis@lunenfeld.ca.

Show MeSH
Related in: MedlinePlus