Details

Autor(en) / Beteiligte

• Cohen, Bruce I.
• Presnell, Scott R.
• Cohen, Fred E.

Titel

[13] Pattern-based approaches to protein structure prediction

Ist Teil von

• Methods in Enzymology, 1991, Vol.202, p.252-268

Ort / Verlag

United States: Elsevier Science & Technology

Erscheinungsjahr

1991

Quelle

MEDLINE

Beschreibungen/Notizen

• This chapter discusses the principles of a set of sequence-structure correlates that can be used to predict secondary structure from protein primary sequence. It highlights an algorithm for finding turns and describes some of the requirements for a pattern language that facilitates the identification of sequence-structure correlates. It also presents the pattern language and examples of patterns that can be used to recognize turns or loops and α helices. The protein domains of known tertiary structure are used as sources for the development of hypothetical sequence-structure correlates and as specimens to examine for potential correlates. To eliminate potential bias, a clear division is made between sequences used to develop sequence-structure correlates and sequences used for testing. Several physical principles are used to guide the development of patterns for locating turns. These include local maxima in hydrophilicity; putative secondary structure identification and avoidance; special backbone dihedral angle restrictions on proline; and weakly hydrophilic regions appropriately spaced from well-defined turns. This also suggests a natural hierarchy of patterns reflecting the relative merits of each concept. Turns form a useful division of the protein sequence into sub-segments that contain no more than one piece of secondary structure. In the case of all helical proteins, it appears possible to recognize helical core sequences consistently. The exact boundaries of the helices have proved more problematic.