A Comparative Analysis of Motif Discovery Algorithms

Abstract

One of the major challenges in bioinformatics is the development of efficient computational algorithms for biological sequence motif discovery. In the post-genomic era, the ability to predict the behavior, the function, or the structure of biological entities or motifs such as genes and proteins, as well as interactions among them, play a fundamental role in the discovery of information to help explain biological mechanisms. This necessitated the development of computational methods for identifying these entities. Consequently, a large number of motif finding algorithms have been implemented and applied to various organisms over the past decade. This paper presents a comparative analysis of the latest developments in motif finding algorithms and proposed an algorithm for motif discovery based on a combinatorial approach of pattern driven and statistical based approach. The proposed algorithm, Suffix Tree Gene Enrichment Motif Searching (STGEMS) as reported in [30], proved effective in identifying motifs from organisms with peculiarity in their genomic structure such as the AT-rich sequence of the malaria parasite, P. falciparum. The empirical time analysis of seven motif discovery algorithms was evaluated using four sets of genes from the intraerythrocytic development cycle of P. falciparum. The result shows that algorithms based on a combinatorial approach are more desirable.

Template-based mapping of dynamic motifs in tissue morphogenesis

“…Relying on our time series representation, these explorations could be done using de-novo motif discovery algorithms, in which a sequence dataset is searched for statistically overrepresented segments in a fast, systematic, and unbiased manner [ 53 , 54 ]. Such modular decomposition approaches proved to be transformative in dealing with large volumes of data from sequencing and structural studies of DNA, RNA, and proteins [ 55 – 57 ]. In the future, this could serve as the basis for recovering a complete dictionary of tissue behaviors.…”

Capturing the geometry, function, and evolution of enzymes with 3D templates

“…These usually consist of residue coordinates (a subset of main or side chain atoms, or average coordinates) and a set of geometrical or physical/chemical constraints to permit tunable matching specificity 4,5 . Special tailor‐made algorithms are used to query protein structures and find template matches, working as a structural equivalent of sequence motif searching methods 6–9 . Templates have been used extensively for the identification of functional sites in protein structures 10 .…”

“… 4 , 5 Special tailor‐made algorithms are used to query protein structures and find template matches, working as a structural equivalent of sequence motif searching methods. 6 , 7 , 8 , 9 Templates have been used extensively for the identification of functional sites in protein structures. 10 More specifically, sites facilitating DNA binding, 11 metal coordination motifs, 12 , 13 catalytic sites, 5 , 14 and ligand binding cavities 15 , 16 have been studied in our group, and template‐matching methods, able to identify such sites, have been incorporated in functional annotation pipelines.…”