Ems1: An Elegant Algorithm for Edit Distance Based Motif Search

Abstract

Motifs are biologically significant patterns found in DNA/protein sequences. Given a set of biological sequences, the problem of identifying the motifs is very challenging. This problem has been well studied in computational biology. Identifying motifs through experimental processes is extremely expensive and time consuming. This is one of the factors influencing computational biologists to come up with novel computational methods to predict motifs. Several motif models have been proposed in the literature and for each model numerous algorithms have been devised. Three popular motif models are (l, d)-motif search or Planted Motif Search (PMS), Simple Motif Search (SMS), and Edit-distance based Motif Search (EMS). For PMS and SMS several algorithms have been proposed and implemented. On the other hand, even though some algorithms exist in the literature for the problem of EMS, no implementations of these algorithms are known. This is mainly because the proposed algorithms are complex. In this paper we present an elegant algorithm for EMS. We have implemented this algorithm and compared it against 14 other algorithms in terms of sensitivity and specificity. Our experimental results indicate that the new algorithm is very competitive in practice.

Efficient sequential and parallel algorithms for finding edit distance based motifs

“…For every ( l,d ) combination we report the average time taken by 4 runs. We compare the following four implementations: EMS1: A modified implementation of the algorithm in [13] which considered the neighborhood of only l -mers whereas the modified version considers the neighborhood of all k -mers where l − d ≤ k ≤ l + d . EMS2: A faster implementation of our sequential algorithm which uses tries for storing candidate motifs where each node of the trie stores an array of pointers to each children of the node. However, this makes the space required to store a tree node dependent on the size of the alphabet Σ .…”

Improved algorithms for finding edit distance based motifs

“…Using equations (I), (2), (3) and (4), Pathak et al [13] gave an algorithm that stores all possible candidate motifs in an array of size 1 �l l . However the algorithm is inefficient in generating the neighborhood as the same candidate motif is generated by several combinations of the basic edit operations.…”

“…Following a useful practice for PMS algorithms, Pathak et al [13] evaluated their algorithm on certain instances that are considered challenging for PMS: (9,2), (11,3), (13,4) and so on [1], and are generated as follows: n = 20 random DNA/protein strings of length m = 600, and a short random string M of length l are generated according to the independent identically distributed (i.i.d) model. A separate random d hamming distance neighbor of M is "planted" in each of the n input strings.…”

“…The challenging instances for EMS turn out to be (8,1), (12,2), (16, 3), (20,4) and so on. To compare with [13], we consider both types of instances, specifically, (8,1), (9,2), (12,2), (11,3), (13,4). …”

“…The algorithms for EMS are customarily evaluated on several challenging instances such as (9,2), (11,3), (13,4), (15,5), and so on. The best previously known algorithm, EMSl, solves up to instance (11,3) in estimated 3 days.…”