Modified circular dichroism data set for validation of secondary structure estimate algorithms

This data set, as an Excel file, was created by simulation of nine far UV CD spectra from the Protein CD Data Base. Spectra were fit as the sum of several (usually four) Gaussian peaks. These noise-free images of the CD spectra were then modified - creating "imperfect data" - by shifting spectral wavelength, intensity and offset, or by varying the noise level, or the ratio of peaks at 192 and 290 nm for a camphor-10-sulphonic acid reference standard. The extent of modification was based on literature data. These modified spectra can help validate the robustness of algorithms used to estimate the secondary structure content of proteins. This was demonstrated for the BeStSel algorithm in the associated publication.
Understanding the robustness of a method to imperfect calibration and spectral factors is critical in the validation of analytical methods using the ICH Q2(R1) protocol, the standard for the pharmaceutical and biopharmaceutical industries.