Editing Protein structure prediction (section)

===Historical perspective===
To date, over 20 different secondary structure prediction methods have been developed. One of the first algorithms was [[Chou–Fasman method]], which relies predominantly on probability parameters determined from relative frequencies of each amino acid's appearance in each type of secondary structure.<ref name="Chou">{{cite journal |vauthors=Chou PY, Fasman GD |title=Prediction of protein conformation |journal=Biochemistry |volume=13 |issue=2 |pages=222–45 |date=January 1974 |pmid=4358940 |doi=10.1021/bi00699a002}}</ref> The original Chou-Fasman parameters, determined from the small sample of structures solved in the mid-1970s, produce poor results compared to modern methods, though the parameterization has been updated since it was first published. The Chou-Fasman method is roughly 50–60% accurate in predicting secondary structures.<ref name="Mount"/>

The next notable program was the [[GOR method]] is an [[information theory]]-based method. It uses the more powerful probabilistic technique of [[Bayesian inference]].<ref name="Garnier">{{cite journal |vauthors=Garnier J, Osguthorpe DJ, Robson B |title=Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins |journal=Journal of Molecular Biology |volume=120 |issue=1 |pages=97–120 |date=March 1978 |pmid=642007 |doi=10.1016/0022-2836(78)90297-8}}</ref> The GOR method takes into account not only the probability of each amino acid having a particular secondary structure, but also the [[conditional probability]] of the amino acid assuming each structure given the contributions of its neighbors (it does not assume that the neighbors have that same structure). The approach is both more sensitive and more accurate than that of Chou and Fasman because amino acid structural propensities are only strong for a small number of amino acids such as [[proline]] and [[glycine]]. Weak contributions from each of many neighbors can add up to strong effects overall. The original GOR method was roughly 65% accurate and is dramatically more successful in predicting alpha helices than beta sheets, which it frequently mispredicted as loops or disorganized regions.<ref name="Mount"/>

Another big step forward, was using [[machine learning]] methods. First [[artificial neural network]]s methods were used. As a training sets they use solved structures to identify common sequence motifs associated with particular arrangements of secondary structures. These methods are over 70% accurate in their predictions, although beta strands are still often underpredicted due to the lack of three-dimensional structural information that would allow assessment of [[hydrogen bonding]] patterns that can promote formation of the extended conformation required for the presence of a complete beta sheet.<ref name="Mount"/> [[Psipred|PSIPRED]] and [[Jpred|JPRED]] are some of the most known programs based on neural networks for protein secondary structure prediction. Next, [[support vector machine]]s have proven particularly useful for predicting the locations of [[turn (biochemistry)|turns]], which are difficult to identify with statistical methods.<ref name="Pham">{{cite journal |vauthors=Pham TH, Satou K, Ho TB |title=Support vector machines for prediction and analysis of beta and gamma-turns in proteins |journal=Journal of Bioinformatics and Computational Biology |volume=3 |issue=2 |pages=343–58 |date=April 2005 |pmid=15852509 |doi=10.1142/S0219720005001089}}</ref><ref name="Zhang">{{cite journal |vauthors=Zhang Q, Yoon S, Welsh WJ |title=Improved method for predicting beta-turn using support vector machine |journal=Bioinformatics |volume=21 |issue=10 |pages=2370–4 |date=May 2005 |pmid=15797917 |doi=10.1093/bioinformatics/bti358 |doi-access=}}</ref>

Extensions of machine learning techniques attempt to predict more fine-grained local properties of proteins, such as [[protein backbone|backbone]] [[dihedral angle]]s in unassigned regions. Both SVMs<ref name="Zimmermann">{{cite journal |vauthors=Zimmermann O, Hansmann UH |title=Support vector machines for prediction of dihedral angle regions |journal=Bioinformatics |volume=22 |issue=24 |pages=3009–15 |date=December 2006 |pmid=17005536 |doi=10.1093/bioinformatics/btl489 |doi-access=}}</ref> and neural networks<ref name="Kuang">{{cite journal |vauthors=Kuang R, Leslie CS, Yang AS |title=Protein backbone angle prediction with machine learning approaches |journal=Bioinformatics |volume=20 |issue=10 |pages=1612–21 |date=July 2004 |pmid=14988121 |doi=10.1093/bioinformatics/bth136 |doi-access=free}}</ref> have been applied to this problem.<ref name="Pham"/> More recently, real-value torsion angles can be accurately predicted by SPINE-X and successfully employed for ab initio structure prediction.<ref name="torsion">{{cite journal |vauthors=Faraggi E, Yang Y, Zhang S, Zhou Y |title=Predicting continuous local structure and the effect of its substitution for secondary structure in fragment-free protein structure prediction |journal=Structure |volume=17 |issue=11 |pages=1515–27 |date=November 2009 |pmid=19913486 |pmc=2778607 |doi=10.1016/j.str.2009.09.006}}</ref>