Computing Reviews

ARIADNE: pattern-directed inference and hierarchical abstraction in protein structure recognition
Lathrop R., Webster T., Smith T. Communications of the ACM30(11):909-921,1987.Type:Article
Date Reviewed: 09/01/88

ARIADNE, the Cretan princess who helped Theseus find his way out of the labyrinth, is the namesake of a hierarchical pattern-directed inference system for the ill-structured problem area of protein structure analysis. ARIADNE identifies the optimal match between a given pattern descriptor and a protein sequence by abstracting intermediate levels of structural organization.

The major limitation of current biosequence comparative methods is that they compare primary sequences and therefore require substantial primary sequence similarity in order to make inferences about protein structure. The microworld is characterized by recognizable higher orders of organization obscured by a high degree of uncertainty and imprecision, and the general approach should be applicable to other similar ill-structured problem areas. ARIADNE facilitates direct expression and manipulation of higher-order structures. This allows direct use of secondary structure predictions that in turn permits a search for similarities at a higher level than that of the primary sequence. The system is implemented in Lisp on a Symbolics 3600 and is useful for continuing research in the fields of both molecular biology and machine learning. ARIADNE has been used to find a novel proposed alignment of the aminoacyl-tRNA synthetases.

Reviewer:  Razvan Andonie Review #: CR112370

Reproduction in whole or in part without permission is prohibited.   Copyright 2024 ComputingReviews.com™
Terms of Use
| Privacy Policy