Recognition and Clarification of Cross Peaks in Crowded Two-Dimensional Correlation Spectra

Abstract

A new scheme is suggested for the automated recognition of cross peaks in two-dimensional correlation spectra. It makes use of prior knowledge of the chemical shifts derived from a symmetry-based search of the two-dimensionalJspectrum, using this information to construct a two-dimensional “chemical-shift grid” that may be overlaid on the corresponding correlation spectrum (COSY, TOCSY, EXSY, or NOESY). The search for local symmetry (or antisymmetry) in the correlation spectrum may then be restricted to intersections on the chemical-shift grid instead of a point-by-point examination of the entire two-dimensional array. This greatly accelerates the search program. Applied to the 400 MHz proton correlation (DQ-COSY) spectrum of 1-dehydrotestosterone, this automated routine locates 28 pairs of cross peaks, and in particular, identifies a tight cluster ofeightinterpenetrating spin multiplets. This latter finding is confirmed by a different method that reconstructs individual cross peaks from partially exposed segments and then strips them away one at a time.