Computer Science and Engineering, Department of


First Advisor

Jitender S. Deogun

Date of this Version

Fall 12-4-2020

Document Type



A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Doctor of Philosophy, Major: Computer Science, Under the Supervision of Professor Jitender S. Deogun. Lincoln, Nebraska: December, 2020

Copyright © 2020 Sairam Behera


In this dissertation, we worked on several algorithmic problems in bioinformatics using mainly three approaches: (a) a streaming model, (b) sux-tree based indexing, and (c) minwise-hashing (minhash) and locality-sensitive hashing (LSH). The streaming models are useful for large data problems where a good approximation needs to be achieved with limited space usage. We developed an approximation algorithm (Kmer-Estimate) using the streaming approach to obtain a better estimation of the frequency of k-mer counts. A k-mer, a subsequence of length k, plays an important role in many bioinformatics analyses such as genome distance estimation. We also developed new methods that use sux tree, a trie data structure, for alignment-free, non-pairwise algorithms for a conserved non-coding sequence (CNS) identification problem. We provided two different algorithms: STAG-CNS to identify exact-matched CNSs and DiCE to identify CNSs with mismatches. Using our algorithms, CNSs among various grass species were identified. A different approach was employed for identification of longer CNSs ( 100 bp, mostly found in animals). In our new method (MinCNE), the minhash approach was used to estimate the Jaccard similarity. Using also LSH, k-mers extracted from genomic sequences were clustered and CNSs were identified. Another new algorithm (MinIsoClust) that also uses minhash and LSH techniques was developed for an isoform clustering problem. Isoforms are generated from the same gene but by alternative splicing. As the isoform sequences share some exons but in different combinations, regular sequencing clustering methods do not work well. Our algorithm generates clusters for isoform sequences based on their shared minhash signatures. Finally, we discuss de novo transcriptome assembly algorithms and how to improve the assembly accuracy using ensemble approaches. First, we did a comprehensive performance analysis on different transcriptome assemblers using simulated benchmark datasets. Then, we developed a new ensemble approach (Minsemble) for the de novo transcriptome assembly problem that integrates isoform-clustering using minhash technique to identify potentially correct transcripts from various de novo transcriptome assemblers. Minsemble identified more correctly assembled transcripts as well as genes compared to other de novo and ensemble methods.

Adviser: Jitender S. Deogun