Peptide Detectability Predictor


dqmodel

DQmodel (Detectability and Quantity model) is a software for fitting artificial neural network models for peptide detectability, which is a probabilistic measure of the sequence preferences of identified peptide in Shotgun proteomics experiments. The model considers protein quantity and thus can model effective detectability and improve standard detectability prediction. From machine learning point of views, the model explicitly estimates the values for the hidden variables - protein quantities, each of which is shared by a subset of instances - peptides from each protein - in the training dataset. From proteomics point of view, effective detectability which combines standard detectability and protein quantity is a more sample specific measure compared with standard detectability or the concept of proteotypic peptide.

DQmodel currently has the following functionalities:

  • Given a confident set of identified proteins and the peptide identifications from them, the program fits a modular artificial neural network model for peptide detectability.

  • If a pre-trained model is available, the program predicts standard detectability for all tryptic peptides from all proteins in a given proteome.

  • The program can perform protein inference and a peptide-count like label free quantification after predicting peptide detectabilities. That requires the  Bayesian protein inference program MSBayesPro.

  • It computes standard protein detectabilities (and effective protein detectabilities if protein quantity is positive) based on predicted peptide detectabilities.

  • Visualize peptide detectability profile for one or a few proteins.

A previous version of peptide detectability predictor is available here, its prediction is based on a hard-wired pre-trained model trained on a sample of 12 standard proteins at equal abundances.

Prerequisites

  • MATLAB version 7.9 or later

  • Python or Perl (or Active Perl for Windows); using Python gives DQmodel better speed through the SuffixTree package.  SuffixTree-0.7.1, can be obtained from here.

Downloads

DQmodel code
License
Sample datasets


References

Yong Fuga Li, Randy J. Arnold, Haixu Tang, and Predrag Radivojac (2010). The importance of peptide detectability for protein identification, quantification, and experiment design in MS/MS proteomics. Journal of Proteome Research.  9(12): 6288–6297. PDF

Yong Fuga Li, Randy Arnold, Yixue Li, Predrag Radivojac, Quanhu Sheng and Haixu Tang. A Bayesian approach to protein inference problem in shotgun proteomics. J Comput Biol (2009) 16(8): 1-11. PDF

Yong Fuga Li, Randy Arnold, Yixue Li, Predrag Radivojac, Quanhu Sheng and Haixu Tang. A Bayesian approach to protein inference problem in shotgun proteomics. RECOMB 2008; & LNBI 4955, pp. 167 - 180, 2008. PDF

H. Tang, R. J. Arnold, P. Alves, Z. Xun, D. E. Clemmer, M. V. Novotny, J. P. Reilly and P. Radivojac, A computational approach toward label-free protein quantification using predicted peptide detectability. ISMB (Supplement of Bioinformatics) 2006: 481-488

 


 

               Last modified: 09/07/2010, Questions?