Deconvolution of Complex 1D NMR Spectra Using Objective Model Selection

摘要: Fluorine (19F) NMR has emerged as a useful tool for characterization of slow dynamics in 19F-labeled proteins. One-dimensional (1D) 19F spectra proteins can be broad, irregular and complex, due to exchange probe nuclei between distinct electrostatic environments; therefore cannot deconvoluted analyzed an objective way using currently available software. We have developed Python-based deconvolution program, decon1d, which uses Bayesian information criteria (BIC) objectively determine model (number peaks) would most likely produce the experimentally obtained data. The method also allows fitting intermediate spectra, is not supported by current software absence specific kinetic model. In methods, determination best data done manually through comparison residual error values, time consuming requires selection user. contrast, BIC used decond1d provides quantitative that penalizes complexity helping prevent over-fitting identification parsimonious decon1d program freely downloadable Python script at project website (https://github.com/hughests/decon1d/).