Authors:
Thanapat Kangkachit
and
Kitsana Waiyamai
Affiliation:
Kasetsart University, Thailand
Keyword(s):
Complete substitution group, Fuzzy concept lattice, Reactive motifs, Enzyme function classification, Binding and catalytic site, Amino acid sustitution matrix, Biochemical knowledge.
Related
Ontology
Subjects/Areas/Topics:
Bioinformatics
;
Biomedical Engineering
;
Data Mining and Machine Learning
;
Sequence Analysis
;
Structure Prediction
Abstract:
Reactive motifs are short conserved regions discovered from binding and catalytic sites of enzymes sequences.
Thus, reactive motifs provide more biological meaning than statistic-based motifs because they are directly
extracted from where the chemical reaction mechanism occurs. Main problem of discovering reactive motifs
is that only 4.94% enzymes sequences contain sites information. To overcome this problem, we present fuzzy
concept lattice-based (FCL-based) method for discovering more general reactive motifs by incorporating biochemical
knowledge. Fuzzy concept lattices are used to represent both binary and multi-value biochemical
knowledge. The fuzzy concept lattice Join operator is applied to determine complete substitution groups that
obtains more general reactive motifs. Experiments are conducted among different methods of determining
complete substitution groups: FCL-based, concecpt lattice-based (CL-based) and similarity-based method.
Experimental results show that FCL-based
method significantly outperforms other methods in term of coverage
value and F-measure with SVM learning algorithm. Therefore, fuzzy concept lattice provides more
efficient computational support for complete substitution groups operation than that of other existing methods.
(More)