Rasiah Loganantharaj

Identification of SSR markers associated with height using pool-based genome-wide association mapping in sorghum

Sorghum has been proposed as a potential energy crop. However, it has been traditionally bred for grain yield and forage quality, not traits related to bioenergy production. To develop tools for genetic improvement of bioenergy-related traits such as height, genetic markers associated with these traits have first to be identified. Association mapping has been extensively used in humans and in some crop plants for this purpose. However, genome-wide association mapping using the whole association panel is costly and time-consuming. A variation of this method called pool-based genome-wide association mapping has been extensively used in humans. In this variation, pools of individuals with contrasting phenotypes, instead of the whole panel, are screened with genetic markers and polymorphic markers are confirmed by screening the individuals in the pools. Here, we identified several new simple sequence repeats (SSR) markers associated with height using this pool-based genome-wide association mapping in sorghum. After screening the tall and short pools of sorghum accessions from the sorghum Mini Core collection developed at the International Crops Research Institute for the Semi-Arid Tropics with 703 SSR markers, we have identified four markers that are closely associated with sorghum height on chromosomes 2, 6, and 9. Comparison with published maps indicates that all four markers are clustered with markers previously mapped to height or height-related traits and with candidate genes involved in regulating plant height such as FtsZ, Ugt, and GA 2-oxidase. The mapping method can be applied to other crop plants for which a high-throughput genome-wide association mapping platform is not yet available.

PAPST, a User Friendly and Powerful Java Platform for ChIP-Seq Peak Co-Localization Analysis and Beyond

Comparative co-localization analysis of transcription factors (TFs) and epigenetic marks (EMs) in specific biological contexts is one of the most critical areas of ChIP-Seq data analysis beyond peak calling. Yet there is a significant lack of user-friendly and powerful tools geared towards co-localization analysis based exploratory research. Most tools currently used for co-localization analysis are command line only and require extensive installation procedures and Linux expertise. Online tools partially address the usability issues of command line tools, but slow response times and few customization features make them unsuitable for rapid data-driven interactive exploratory research. We have developed PAPST: Peak Assignment and Profile Search Tool, a user-friendly yet powerful platform with a unique design, which integrates both gene-centric and peak-centric co-localization analysis into a single package. Most of PAPST’s functions can be completed in less than five seconds, allowing quick cycles of data-driven hypothesis generation and testing. With PAPST, a researcher with or without computational expertise can perform sophisticated co-localization pattern analysis of multiple TFs and EMs, either against all known genes or a set of genomic regions obtained from public repositories or prior analysis. PAPST is a versatile, efficient, and customizable tool for genome-wide data-driven exploratory research. Creatively used, PAPST can be quickly applied to any genomic data analysis that involves a comparison of two or more sets of genomic coordinate intervals, making it a powerful tool for a wide range of exploratory genomic research. We first present PAPST’s general purpose features then apply it to several public ChIP-Seq data sets to demonstrate its rapid execution and potential for cutting-edge research with a case study in enhancer analysis. To our knowledge, PAPST is the first software of its kind to provide efficient and sophisticated post peak-calling ChIP-Seq data analysis as an easy-to-use interactive application. PAPST is available at https://github.com/paulbible/papst and is a public domain work.

Reasoning about networks of temporal relations and its applications to problem solving

An interval algebra (IA) has been proposed as a model for representing and reasoning about qualitative temporal relations between time intervals. Unfortunately, reasoning tasks with IA that involve deciding the satisfiability of the temporal constraints, or providing all the satisfying instances of the temporal constraints, areNP-complete. That is, solving these problems are computationally exponential in the “worst case.” However, several directions in improving their computational performance are still possible. This paper presents a new backtracking algorithm for finding a solution called consistent scenario. This algorithm has anO(n3) best-case complexity, compared toO(n4) of previous known backtrack algorithms, wheren denotes the number of intervals. By computational experiments, we tested the performance of different backtrack algorithms on a set of randomly generated networks with the results favoring our proposal. In this paper, we also present a new path consistency algorithm, which has been used for finding approximate solutions towards the minimal labeling networks. The worst-case complexity of the proposed algorithm is stillO(n3); however, we are able to improve its performance by eliminating the unnecessary duplicate computation as presented in Allens original algorithm, and by employing a most-constrained first principle, which ensures a faster convergence. The performance of the proposed scheme is evaluated through a large set of experimental data.

Parallel path-consistency algorithms for constraint satisfaction

The authors previously proposed (1991) a O(n/sup 3/) path-consistency algorithm which requires O(n/sup 2/) space, whereas other known O(n/sup 3/) time complexity algorithms need O(n/sup 3/) space. They use this algorithm as the main framework for a parallel version.<<ETX>>

The Limitations of Existing Approaches in Improving MicroRNA Target Prediction Accuracy

MicroRNAs (miRNAs) are small (18-24 nt) endogenous RNAs found across diverse phyla involved in posttranscriptional regulation, primarily downregulation of mRNAs. Experimentally determining miRNA-mRNA interactions can be expensive and time-consuming, making the accurate computational prediction of miRNA targets a high priority. Since miRNA-mRNA base pairing in mammals is not perfectly complementary and only a fraction of the identified motifs are real binding sites, accurately predicting miRNA targets remains challenging. The limitations and bottlenecks of existing algorithms and approaches are discussed in this chapter.A new miRNA-mRNA interaction algorithm was implemented in Python (TargetFind) to capture three different modes of association and to maximize detection sensitivity to around 95% for mouse (mm9) and human (hg19) reference data. For human (hg19) data, the prediction accuracy with any one feature among evolutionarily conserved score, multiple targets in a UTR or changes in free energy varied within a close range from 63.5% to 66%. When the results of these features are combined with majority voting, the expected prediction accuracy increases to 69.5%. When all three features are used together, the average best prediction accuracy with tenfold cross validation from the classifiers naïve Bayes, support vector machine, artificial neural network, and decision tree were, respectively, 66.5%, 67.1%, 69%, and 68.4%. The results reveal the advantages and limitations of these approaches.When comparing different sets of features on their strength in predicting true hg19 targets, evolutionarily conserved score slightly outperformed all other features based on thermostability, and target multiplicity. The sophisticated supervised learning algorithms did not improve the prediction accuracy significantly compared to a simple threshold based approach on conservation score or combining the results of each feature with majority agreements. The targets from randomly generated UTRs behaved similar to that of noninteracting pairs with respect to changes in free energy. Availability of additional experimental data describing noninteracting pairs will advance our understanding of the characteristics and the factors positively and negatively influencing these interactions.

Towards recognition of protein function based on its structure using deep convolutional networks

This paper proposes a novel method for protein function recognition using deep learning. Recently, deep convolutional neural networks (DCNNs) demonstrated high performances in many areas of pattern recognition. Protein function is often associated with its tertiary structure denoting the active domain of a protein. This investigation develops a novel DCNN for protein functionality recognition based on its tertiary structure. Two rounds of experiments are performed. The initial experiment on tertiary protein structure alignment shows promising performances (94% accuracy rate) such that it shows the model robustness against rotations, local translations, and scales of the 3D structure. With these results, the main experiments contain five different datasets obtained by similarity measures between pairs of gene ontology terms. The experimental results for protein function recognition on selected datasets show 87.6% and 80.7% maximum and average accuracy rates respectively. The initial success of the DCNN in tertiary protein structure recognition supports further investigations with respect to tertiary protein retrieval and pattern mining on large scale problems.

Experimenting with a temporal constraint propagation algorithm

Abstract3-consistency algorithm for temporal constraint propagation over interval-based network, proposed by James Allen, is finding its use in many practical temporal reasoning systems. Apart from the polynomial behavior of this algorithm with respect to the number of nodes in the network, very little is known about its time complexity with respect to other properties of the initially given temporal constraints. In this article we have reported some of our results analyzing the complexity with respect to some structural parameters of the input constraint network. We have identified some regions, with respect to the structural parameters of the input network, where the algorithm takes much more time than it needs over other regions. Similar features have been observed in recent studies on NP-hard problems. Average case complexity of Allens algorithm is also studied empirically, over a hundred thousand randomly generated networks, and the growth rate is observed to be of the order of quadratic with respect to the problem size (at least up to node 40, and expected to be lower above that). We have analyzed our data statistically to develop a model with which one can calculate the expected time to be consumed by the algorithm for a given input network.

The effects of shared information on semantic calculations in the gene ontology

The structured vocabulary that describes gene function, the gene ontology (GO), serves as a powerful tool in biological research. One application of GO in computational biology calculates semantic similarity between two concepts to make inferences about the functional similarity of genes. A class of term similarity algorithms explicitly calculates the shared information (SI) between concepts then substitutes this calculation into traditional term similarity measures such as Resnik, Lin, and Jiang-Conrath. Alternative SI approaches, when combined with ontology choice and term similarity type, lead to many gene-to-gene similarity measures. No thorough investigation has been made into the behavior, complexity, and performance of semantic methods derived from distinct SI approaches. We apply bootstrapping to compare the generalized performance of 57 gene-to-gene semantic measures across six benchmarks. Considering the number of measures, we additionally evaluate whether these methods can be leveraged through ensemble machine learning to improve prediction performance. Results showed that the choice of ontology type most strongly influenced performance across all evaluations. Combining measures into an ensemble classifier reduces cross-validation error beyond any individual measure for protein interaction prediction. This improvement resulted from information gained through the combination of ontology types as ensemble methods within each GO type offered no improvement. These results demonstrate that multiple SI measures can be leveraged for machine learning tasks such as automated gene function prediction by incorporating methods from across the ontologies. To facilitate future research in this area, we developed the GO Graph Tool Kit (GGTK), an open source C++ library with Python interface (github.com/paulbible/ggtk).

Extension of Petri Nets for Representing and Reasoning with Tasks with Imprecise Durations

This paper presents an extension of Petri net framework with imprecise temporal properties. We use possibility theory to represent imprecise time by time-stamping tokens and assigning durations to firing of the transitions. A method for approximation of an arbitrary temporal distribution with a set of possibilistic intervals is used to introduce the composition operation for two possibilistic temporal distributions. We developed a method to determining an effective enabling time of a transition with incoming tokens with possibilistic distributions. The utility of the proposed theory is illustrated using an example of an automated manufacturing system. The proposed approach is novel and has a broad utility beyond a timed Petri network and its applications.

An overview of a synergetic combination of local search with evolutionary learning to solve optimization problems

We describe a method for solving combinatorial optimization problem that combines best aspects of local search and genetic algorithms. We formulate combinatorial optimization problems as state space search problems. While local search methods, such as hill climbing, are computationally efficient, they suffers from local minima traps. Global search methods are guaranteed to find optimal solutions, but are not always feasible. We favor a polynomial time technique that delivers solutions closer to optimal by modifying the search space of the local search method. We demonstrate our strategy on a single-machine scheduling problem with two objective functions: (1) minimizing average job completion time, and (2) minimizing total tardiness. We apply the technique to optimally schedule the robot arm of an automated retrieval system. Obtaining optimal solutions to such scheduling problems is computationally intractable, but experimental results show our technique produces better solutions than those found by genetic algorithm with random key encoding.