|
Abstract:
|
Proteins are signi cant biological macromolecules of polymeric nature
(polypeptides), which contain amino acids and are basic structural units of each cell.
Their contents include 20+3 amino acids and, as a consequence, they are presented
in biological databases as sequences formed from 23 di erent characters. Proteins
can be classi ed based on their primary structure, secondary structure, function etc.
One of possible classi cations of proteins by their function is related to their contents
in a certain cluster of ortholologous groups (COGs). This classi cation is based on
the previous comparison of proteins by their similarities in their primary structures,
which is most often a result of homology, i.e. their mutual (evolutionary) origin.
COG database is obtained by comparison of the known and predicted proteins encoded
in the completely sequenced prokaryotic (archaea and bacteria) genomes and
their classi cation by orthology. The proteins are classi ed in 25 categories which
can be ordered in three basic functional groups (the proteins responsible for: (1)
information storage and processing; (2) cellular processes and signaling; and (3)
metabolism), or in a group of poorly characterized proteins. Classi cation of proteins
by their contents in certain COG category (euKaryote Orthologous Groups-
KOG for eukaryotic organisms) is signi cant for better understanding of biological
processes and various pathological conditions in people and other organisms.
The dissertation proposed the model for classi cation of proteins in COG categories
based on amino acid n-grams (sequences of n- length). The set of data contains
protein sequences of genomes from 8 di erent taxonomic classes [TKL97] of bacteria
(Aqui cales, Bacteroidia, Chlamydiales, Chlorobia, Chloro exia, Cytophagia,
Deinococci, Prochlorales), which are known to have been classi ed by COG categories.
The new method is presented, based on the generalized systems of Boolean
equations, used for separation of n-grams characteristic for proteins of corresponding
COG categories. The presented method signi cantly reduces the number of
processed n-grams in comparison to previously used methods of n-gram analysis,
iv
thus more memory space is provided and less time for protein procession is necessary.
The previously known methods for classi cation of proteins by functional categories
compared each new protein (whose function had to be determined) to the set of all
proteins which had already been classi ed by functions in order to determine the
group which contained most similar proteins to the one which was to be classi ed.
In relation to the previous, the advantage of the new method is in its avoidance
of sequence-sequence comparison and in search for those patterns (n-grams, up to
10 long) in a protein which are characteristic of the corresponding COG category.
The selected patterns are added to a corresponding COG category and describe
sequences of certain length, which have previously appeared in that COG category
only, not in the proteins of other COG categories.
On the basis of the proposed method, the predictor for determination of the corresponding
COG category for a new protein is implemented. Minimal precision of the
prediction is one of the predictors arguments. During the test phase the constructed
predictor shown excellent results, with maximal precision of 99% reached for some
proteins.
According to its properties and relatively simple construction, the proposed method
can be applied in similar domains where the solution of problem is based on n-gram
sequence analysis. |