Wisnu Ananta Kusuma

Tandem repeats analysis in DNA sequences based on improved Burrows-Wheeler transform

The enormous amount of short reads generated by the new DNA sequencing technologies call for the development of fast and accurate read alignment programs. A first generation of hash table-based methods has been developed, including Mapping and Assembly with Quality (MAQ), which is accurate, feature rich and fast enough to align short reads from a single individual. However, MAQ does not support gapped alignment for single-end reads, which makes it unsuitable for alignment of longer reads where indels may occur frequently. The speed of MAQ is also a concern when the alignment is scaled up to the resequencing of hundreds of individuals. Therefore, we carried out an in-depth performance analysis of BWA a popular BWT-based aligner and discovered that its performance is significantly better than MAQ although, it has drawbacks regarding execution speed, time complexity and accuracy. Based on those factors we implemented an improved Burrows-Wheeler Alignment algorithm (BWA), anew read alignment package which is original BWT optimized by source code of Ziv-Lempel (LZ-77) sliding window technique and prefix trie string matching, to efficiently search for inexact and exact matches on tandem repeats against a large reference sequence genome. Our analysis show that search speed of improved BWA significantly increased by approximately 1.40 ×faster than MAQ-32 while achieving sufficiently higher accuracy with percent confidence of 96.7 % and 93.0 %. Moreover, it is more efficient to search exact and inexact matches supported by percent error of 0.05 % single ends and 0.04 % for paired end reads also more effective to search for left and right overlap tandem repeat at percent confidence of 88.9%.

Learning similarity functions for binary strings via genetic programming

Data that encode the presence of some characteristics typically can be represented as binary strings. We need similarity functions for binary strings in order to classify or cluster them. Existing similarity functions, however, do not take advantage of training data, which are often available. We believe that similarity functions should be data-specific. To this end, we use genetic programming (GP) to learn similarity functions from training data. We propose a novel fitness function that considers five aspects of good similarity functions, i.e. recall, magnitude, zero-division, identity and symmetry. We also report mostly-used math operators from extensive literature review. Experiment results show that GP-based similarity functions outperform the well-known Tanimoto function in most datasets in terms of classification accuracy using SVMs. In addition, those GP-based similarity functions are simpler: using fewer numbers of operators and operands. This suggests that our proposed fitness function for GP is justifiable for learning similarity functions.

A classification system for jamu efficacy based on formula using support vector machine and k-means algorithm as a feature selection

Jamu is an Indonesia herbal medicine made from natural materials such as roots, leaves, fruits, and animals. The purpose of this research is to develop a classification system for jamu efficacy based on the composition of plants using Support Vector Machine (SVM) and to implement the k-means clustering algorithm as a feature selection method. The result of this study was compared to the previous research that using SVM method without feature selection. This study used variances to evaluate the results of clustering. The total of 3138 data herbs and 465 plant species were grouped into 100 clusters with the variance of 0.0094. The managed group succesfully reduced the data dimension into 3047 of jamu sample and 236 species of herbs and plants as features. The result of SVM classification using feature selection yielded the accuracy of 71.5%.

Identification of single nucleotide polymorphism using support vector machine on imbalanced data

The advance of DNA sequencing technology presents a significant bioinformatic challenges in a downstream analysis such as identification of single nucleotide polymorphism (SNP). SNP is the most abundant form of genetic marker and have been one of the most crucial researches in bioinformatics. SNP has been applied in wide area, but analysis of SNP in plants is very limited, as in cultivated soybean (Glycine max L.). This paper discusses the identification of SNP in cultivated soybean using Support Vector Machine (SVM). SVM is trained using positive and negative SNP. Previously, we performed a balancing positive and negative SNP with undersampling and oversampling to obtain training data. As a result, the model which is trained with balanced data has better performance than that with imbalanced data.

Application of decision tree classifier for single nucleotide polymorphism discovery from next-generation sequencing data

Single Nucleotide Polymorphism (SNP) is the most abundant form of genetic variation and proven to be advantageous in diverse genetic-related studies. However, accurate determination of true SNPs from next-generation sequencing (NGS) data is a challenging task due to high error rates of NGS. To overcome this problem, we applied a machine learning method using C4.5 decision tree algorithm to discover SNPs from whole-genome NGS data. In addition, we conducted random undersampling to deal with the imbalanced data. The result shows that the proposed method is able to identify most of the true SNPs with more than 90% recall, but still suffers from a high rate of false-positives.

Parallelization of star alignment

Pair-wise sequence alignment is a technique of comparing the similarity of two organisms. It is the basic technique in DNA sequence alignment. There is an extraordinary number of data sequences when they are compared. Problems when comparing the huge data sequences are accuracy and efficiency. These parameters are contradiction which means reaching faster speed will decrease accuracy, and vice versa. Older method such as Needleman-Wunchs has extremely high computational complexity of O(n2). Similarly, multiple sequence alignment that processes the sequences one by one, called Star Alignment, takes time until O(k2n2). Therefore, they have a timing issue problem while processing the data. However, the computation result still has high accuracy. Consequently, it is very important to get a better way to improve the performance. One of the methods for increasing the speed is parallel computation by using multiple computers work together as a system. This research focuses on finding a faster method to process multiple sequence alignment using the Star method with a parallel computer. Our proposed method is implemented using Message Passing Interface (MPI). The results show that the paralellization of the Star Alignment increased speed up 4-6 times compared to that of using single CPU.

DNA sequencing error correction using spectral alignment

The second generation DNA sequencing technology can generate large number of DNA fragments/reads in a relatively short time. A DNA sequence assembly step is required to obtain whole genome sequences from reads. The assembly process generally uses graph based approach. This approach is very sensitive due to DNA sequencing errors. To obtain the optimal results in assembly process, the error correction step can be performed before or after the assembly process. In this research, we developed a software prototype for correcting DNA sequencing error. We employed the spectral alignment technique implemented as a pre-processing step before the DNA sequence assembly process. We tested our method by using simulated DNA reads containing errors. We measured the results by evaluating the number of nodes. The evaluation results showed that our method can reduce the complexity of graph shown by the decreasing of number of nodes. It can be stated that our method has successfully corrected DNA reads which contain sequencing errors.

Clustering metagenome fragments using growing self organizing map

The microorganism samples taken directly from environment are not easy to assemble because they contains mixtures of microorganism. If sample complexity is very high and comes from highly diverse environment, the difficulty of assembling DNA sequences is increasing since the interspecies chimeras can happen. To avoid this problem, in this research, we proposed binning based on composition using unsupervised learning. We employed trinucleotide and tetranucleotide frequency as features and GSOM algorithm as clustering method. GSOM was implemented to map features into high dimension feature space. We tested our method using small microbial community dataset. The quality of cluster was evaluated based on the following parameters : topographic error, quantization error, and error percentage. The evaluation results show that the best cluster can be obtained using GSOM and tetranucleotide.

A classification system for Jamu efficacy based on formula using Support Vector Machine

Jamu is made from natural materials such as roots, leaves, timber and fruits. Jamu has many variations of formula. The composition of Jamu formula is usually based on empirical data or personal experiences. Thus, the classification for the efficacy of Jamu based on its compositions of plants still remains an interesting task. The purpose of this research is to develop a classification system for Jamu effects based on the composition of plants using Support Vector Machine (SVM). This method is compared to those of previous research using Partial Least Squares Discriminant Analysis (PLS-DA). The result shows that the SVM method with Radial Basis Function (RBF) kernel obtains higher accuracy than those that used PLS-DA.