U
Usman Roshan
Researcher at New Jersey Institute of Technology
Publications - 64
Citations - 1120
Usman Roshan is an academic researcher from New Jersey Institute of Technology. The author has contributed to research in topics: Convolutional neural network & Artificial neural network. The author has an hindex of 14, co-authored 62 publications receiving 1007 citations. Previous affiliations of Usman Roshan include University of Texas at Austin & University of North Carolina at Charlotte.
Papers
More filters
Journal ArticleDOI
Probalign: multiple sequence alignment using partition function posterior probabilities
Usman Roshan,Dennis R. Livesay +1 more
TL;DR: The results indicate that Probalign alignments are generally more accurate than other leading multiple sequence alignment methods (i.e. Probcons, MAFFT and MUSCLE) on the BAliBASE 3.0 protein alignment benchmark and outperforms these methods on the HOMSTRAD and OXBENCH benchmarks.
Journal ArticleDOI
Diametrical clustering for identifying anti-correlated gene clusters
TL;DR: A new diametrical clustering algorithm that explicitly identifies anti-correlated clusters of genes and systems whose mRNA expression patterns oppose the yeast ribosome and proteosome are presented, along with evidence for the inverse transcriptional regulation of a number of cellular systems.
Proceedings ArticleDOI
Rec-I-DCM3: a fast algorithmic technique for reconstructing phylogenetic trees
TL;DR: This paper presents a new technique called Recursive-Iterative-DCM3 (Rec-I- DCM3), which belongs to the family of disk-covering methods (DCMs), and tests this new technique on ten large biological datasets and obtained dramatic speedups as well as significant improvements in accuracy.
Journal ArticleDOI
Ranking causal variants and associated regions in genome-wide association studies by the support vector machine and random forest
TL;DR: If the SVM and RF are applied to the top 2r chi-square-ranked SNPs, it is found that both improve the ranks of causal variants and associated regions and achieve higher power on simulated data.
Journal ArticleDOI
Designing fast converging phylogenetic methods
TL;DR: This study establishes that the new methods outperform both neighbor joining and the previous fast converging methods, returning very accurate large trees, when these other methods do poorly.