applied general statistics croxton cowden pdf

Our method requires primitive computational means and a relatively small training set, while still been comparable to previous work. We use three independent networks, one for each structure (alpha, beta and coil) as the first-level processing unit decision upon the chosen structure for each residue is carried out by a second-level, post-processing unit, which utilizes the Chou and Fasman frequency values F α and F β in order to strengthen and/or deplete the probability of the specific structure under investigation. The present study utilizes expert neural networks for the prediction of proteins secondary structure. The justification of the classification of amino acid sidechains into helix favouring, indifferent and helix inhibiting groups has in part (the inhibiting residues) been supported by a study of their influence on b 0. Of 40 proteins studied 35 fitted the correlation well while the presence of a β-form was demonstrated in three of the remaining five.

Allowance for such effects can at present only be made on a qualitative basis but the phenomenon is known from infrared absorption measurements of proteins and model polyamino acids and from other studies which are independent of the b 0-determination. The presence of a β-form or of left-handed helices seems to influence b 0 markedly.

A statistical analysis of the correlation between the content of certain amino acids in 40 proteins and their optical rotatory dispersion parameter b 0 supports previous direct observations of the interactions between hydroxyl groups of serine and threonine and peptide linkages, which cause interference with the formation of α-helices.

Applied general statistics croxton cowden pdf