Identification and application of the concepts important for accurate and reliable protein secondary structure prediction

Abstract

A protein secondary structure prediction method from multiply aligned homologous sequences is presented with an overall per residue three-state accuracy of 70.1 %. There are two aims: to obtain high accuracy by identification of a set of concepts important for prediction followed by use of linear statistics; and to provide insight into the folding process. The important concepts in secondary structure prediction are identified as: residue conformational propensities, sequence edge effects, moments of hydrophobicity, position of insertions and deletions in aligned homologous sequence, moments of conservation, auto-correlation, residue ratios, secondary structure feedback effects, and filtering. Explicit use of edge effects, moments of conservation, and auto-correlation are new to this paper. The relative importance of the concepts used in prediction was analyzed by stepwise addition of information and examination of weights in the discrimination function. The simple and explicit structure of the prediction allows the method to be reimplemented easily. The accuracy of a prediction is predictable a priori. This permits evaluation of the utility of the prediction: 10% of the chains predicted were identified correctly as having a mean accuracy of >80%. Existing high-accuracy prediction methods are "black-box" predictors based on complex nonlinear statistics (e.g., neural networks in PHD: Rost & Sander, 1993a). For medium-to short-length chains ( 2 9 0 residues and <I70 residues), the prediction method is significantly more accurate ( P < 0.01) than the PHD algorithm (probably the most commonly used algorithm). In combination with the PHD, an algorithm is formed that is significantly more accurate than either method, with an estimated overall three-state accuracy of 72.4%, the highest accuracy reported for any prediction method.

A family of putative MSCRAMMs from Enterococcus faecalis

“…Eight algorithms were initially evaluated for predicting the a-helix and b-sheet content of segments of the A-domains of the staphylococcal MSCRAMMs CNA, ClfA and SdrG for which crystal structure data have been reported (Deivanayagam et al, 1999;Ponnuraj et al, 2003;Symersky et al, 1997). The results showed high variation with the algorithms PHD, DSC and SOPM (Geourjon & Deleage, 1994;King & Sternberg, 1996;Rost & Sander, 1993) being most accurate when aligned with corresponding crystallography data. The mean deviation from solved structures for Cna, SdrG and ClfA was 0?4, 2?2 and 2?8 % for a-helices, and 1?9, 0?6 and 3?3 % for b-sheet, respectively.…”

Optimizing Long Intrinsic Disorder Predictors With Protein Evolutionary Information

“…Another approach trains a second-stage neural network to map the raw predictions across neighboring residues to the final prediction for the current position 25,39 . A third approach uses a set of explicit rules to decide the final prediction based on the raw predictions 40 , which requires the proper integration of domain knowledge.…”

Biophysical Analysis of the N-Terminal Domain from the Human Protein Phosphatase 1 Nuclear Targeting Subunit PNUTS Suggests an Extended Transcription Factor TFIIS-Like Fold

“…The server also incorporates secondary structure prediction using the PSIPRED method [21]. In addition, we also used the Network Protein Sequence Analysis secondary structure prediction server (https://npsa-prabi.ibcp.fr) implementing the MLRC [22], DSC [23], and PHD predictive methods [24]. For identification of the putative fold we utilised the intensive search mode of the Phyre2 online server [25].…”