New algorithms for the analysis of mass spectral profiles from amphibian data

Abstract

Algorithms for analysis of amphibian mass spectral peptides profiles based on HPLCESI/MS (high performance liquid chromatography - electrospray ionisation / mass spectrometry) and tandem mass spectrometry have been developed. The core of the research ·has focused upon the characterisation of the amphibian species by their skin peptides mass spectral profiles and the identification of common peptides among different amphibian species from the skin peptides mass spectral profiles and tandem mass spectra.

An algorithm for automatic detection of time windows which contain significant structures in the total ion current is developed. Further analysis within ea.eh time window then establishes the peptides profile of the specimen.

A heuristic deconvolution algorithm has been developed to extract the molecular weight. of each peptide from its mass spectrum containing multiple charge states of the ions. Two main advantages of this algorithm, compared to other deconvolution algori!hms, are that it is insensitive to the assumed maximum charge state, and the relative abundance in deconvolution spectrum reflects the original concentrations of the compounds in the mixture.

A multiple criteria dynamic window-based clustering algorithm is developed and used to cluster the compounds from peptides profiles of different species or specimens \ to identify alike compounds with sufficiently close molecular weight and retention time.

Overlap algorithms are applied to quantitative measuring the similarity of the peptides profiles of different species or specimens and the similarity of tandem mass spectra of peptides.

The results have shown that skin peptides profiles can be a probe of amphibian species and it is a practicable method to identify common peptides with HPLC-ESI/MS spectra followed by tandem mass spectral analysis.

Date of Award	Dec 2002
Original language	English
Supervisor	John G Hughes (Supervisor)