FTIR Imaging of Protein Microarrays for High Throughput Secondary Structure Determination

The paper introduces a new method designed for high-throughput protein structure determination. It is based on spotting proteins as microarrays at density of ca. 2000–4000 samples per cm2 and recording Fourier transform infrared (FTIR) spectra by FTIR imaging. also library, called cSP92, which contains 92 well-characterized proteins. has been to cover well possible the structural space, both in terms secondary structures higher level structures. Ascending stepwise linear regression (ASLR), partial least square (PLS) regression, support vector machine (SVM) have used correlate spectral characteristics features. ASLR generally provides better results than PLS SVM. observation that prediction good microarray reference attenuated total reflection recorded same validates high throughput approach. Repeated double cross-validation shows approach suitable accuracy determination with root mean standard error 4.9 ± 1.1% α-helix, 4.6 0.8% β-sheet, 6.3 2.2% “other” when using ASLR.