Сайт находится в стадии разработки. Просим извинения за временные неудобства.
Пожалуйста, авторизуйтесь
Логин:
Пароль:
   

Забыли свой пароль?

Следуйте на форму для запроса пароля.

После получения контрольной строки следуйте на форму для смены пароля.
Objective
The MSA software was developed to identify proteins from the peptide mass fingerprint data obtained from MALDI-TOF. The most advantageous application of the methods is the analysis of SDS-PAGE spots, where the mixture of the proteins is the common case. MSA is also the method of choice for transmembrane proteins 1D-separations.

Method
Given the peak list, the protein database and the expected mass range MSA performs three stages of analysis. 
  • Prefiltering: from the protein database those proteins are discarded, which have either the non-allowed mass or which match less than 10% of peaks.
  • Assessment: for each protein candidate the likelihood of identification is calculated using different scores: Bayesian, MOWSE, digestion pattern and peptide length distribution.
  • Meta-scoring: the candidate proteins, each equipped with four above-mentioned scores, are projected onto the self-organizing map (SOM). Above the SOM plane the Gaussian Mixture Model is created. From the GMM the unified probability of identification is assessed for each protein.
Results
Comparing the MSA results with MASCOT and ProFound search engines we observed (1) proteins poorly scored by MASCOT and/or ProFound overcome the significance threshold with MSA; (2) MSA detects the proteins which are absent from the MASCOT and/or ProFound hit lists. SDS-PAGE of human microsomes processed by MSA brought 4-times more identified proteins than the 2D-PAGE separation of the same samples. The presence of proteins recognized in the mixture was confirmed by LC-MS/MS.  

Contact:
proteomics@ibmh.msk.su

<<back>>