PubMed Health⌕ Search

Biomedical subjects

Sanguthevar Rajasekaran

Publications and source records attributed to Sanguthevar Rajasekaran.

6 recordsLinked to original sources
PubMed2006

Identifying offline muscle strength profiles sufficient for short-duration FES-LCE exercise: a PAC learning model approach.

UNLABELLED: Functional electrical stimulation-induced leg cycle ergometry (FES-LCE) provides therapeutic exercise for persons with spinal cord injury (SCI). However, there exists no systematic approach to predict whether an individual has sufficient thigh muscle strength necessary for FES-LCE exercise. OBJECTIVE: To develop and test a Probably Approximately Correct (PAC) learning model as a predictor of thigh muscle strengths sufficient for short-duration FES-LCE exercise and compare the model's performance with other well-known statistical methods. METHODS: Six healthy male individuals with SCI, having age (32.0 +/- 12.5 years), height (1.8 +/- 0.04 m), and weight (79.12 +/- 10.76 kg), participated in static and dynamic experiments. During static experiments, absolute crank torque measurements were used to estimate thigh muscle strengths in response to maximum FES intensities of 70 mA, 105 mA, and 140 mA at fixed crank positions on an FES-LCE. During dynamic experiments, changes in power output measurements were used to classify rider performance as 'Fatigue' or 'No Fatigue' during short-duration FES-LCE at maximum stimulation intensities of 70 mA, 105 mA, and 140 mA and flywheel resistance levels of 0/8th, 1/8th, and 2/8th kilopounds. A Probably Approximately Correct (PAC) learning model was developed to classify static offline muscle strength observations with online rider performances. PAC's discriminatory power was compared with logistic regression (LR), Fisher's linear discriminant analysis (LDA), and an artificial neural network (ANN) model. RESULTS: PAC and ANN learning models correctly identified 100% of the training examples. PAC's average performance on the validation set was 93.1%. The ANN and LR performed comparable with 92.8% and 93.1% accuracy, respectively. The LDA method faired well on the validation set at 89.9%. CONCLUSIONS: PAC performed well in identifying muscle strengths associated with the online performance criterion. Although PAC did not perform best during cross-validation, this model has many advantages over the other methods. PAC can adapt to changes in classification schemes and is more amenable to theoretical analyses than the other methods. PAC learning has an intuitive design and may be a practical choice for classifying muscle strength profiles with well-defined performance criteria.

Adult↗
PubMed2006

Minimotif Miner: a tool for investigating protein function.

In addition to large domains, many short motifs mediate functional post-translational modification of proteins as well as protein-protein interactions and protein trafficking functions. We have constructed a motif database comprising 312 unique motifs and a web-based tool for identifying motifs in proteins. Functional motifs predicted by MnM can be ranked by several approaches, and we validated these scores by analyzing thousands of confirmed examples and by confirming prediction of previously unidentified 14-3-3 motifs in EFF-1.

14-3-3 Proteins↗
PubMed2005

High-performance exact algorithms for motif search.

OBJECTIVE: The human genome project has resulted in the generation of voluminous biological data. Novel computational techniques are called for to extract useful information from this data. One such technique is that of finding patterns that are repeated over many sequences (and possibly over many species). In this paper we study the problem of identifying meaningful patterns (i.e., motifs) from biological data, the motif search problem. METHODS: The general version of the motif search problem is NP-hard. Numerous algorithms have been proposed in the literature to solve this problem. Many of these algorithms fall under the category of heuristics. We concentrate on exact algorithms in this paper. In particular, we concentrate on two different versions of the motif search problem and offer exact algorithms for them. RESULTS: In this paper we present algorithms for two versions of the motif search problem. All of our algorithms are elegant and use only such simple data structures as arrays. For the first version of the problem described as Problem 1 in the paper, we present a simple sorting based algorithm, SMS (Simple Motif Search). This algorithm has been coded and experimental results have been obtained. For the second version of the problem (described in the paper as Problem 2), we present two different algorithms--a deterministic algorithm (called DMS) and a randomized algorithm (Monte Carlo algorithm). We also show how these algorithms can be parallelized. CONCLUSIONS: All the algorithms proposed in this paper are improvements over existing algorithms for these versions of motif search in biological sequence data. The algorithms presented have the potential of performing well in practice.

Algorithms↗
PubMed2005

Randomized and parallel algorithms for distance matrix calculations in multiple sequence alignment.

Multiple sequence alignment (MSA) is a vital problem in biology. Optimal alignment of multiple sequences becomes impractical even for a modest number of sequences since the general version of the problem is NP-hard. Because of the high time complexity of traditional MSA algorithms, even today's fast computers are not able to solve the problem for large number of sequences. In this paper we present a randomized algorithm to calculate distance matrices, which is a major step in many multiple sequence alignment algorithms. The basic idea employed is sampling (along the lines of). We also illustrate how to parallelize this algorithm. In Section we introduce the problem of multiple sequence alignments. In Section we provide a discussion on various methods that have been employed in the literature for Multiple Sequence Alignment. In this section we also introduce our new sampling approach. We extend our randomized algorithm to the case of non-uniform length sequences as well. We show that our algorithms are amenable to parallelism in Section. In Section we back up our claim of speedup and accuracy with empirical data and examples. In Section we provide some concluding remarks.

Algorithms↗
PubMed2003

Parallel pattern identification in biological sequences on clusters.

Tandem repeats are ubiquitous sequence features in both prokaryotic and eukaryotic genomes. They are known to cause several inherited neurological diseases in humans. Identifying these patterns is a highly computation-intensive process. Previous parallel implementations use straightforward domain decomposition based on existing sequential algorithms and rely on parallel machines with low-latency interconnection network and fast hardware support for processor synchronization. Our research exploits the superior cost effectiveness and flexibility achieved through low-cost clusters to speed up biological computations by designing communication-efficient parallel algorithms for pattern identification. This paper presents a low communication-overhead parallel algorithm for pattern identification in biological sequences. Given a biological sequence of length n and a pattern of length m, we conclude an algorithm with five computation/communication phases, each requiring O(n) computation time and only O(p) message units. The low communication overhead of the algorithm is essential in achieving reasonable speedups on clusters, where the inter-processor communication latency is usually higher.

Algorithms↗