Department of Computer Science, Limited Term Appointment/Tenure Track Appointment Positions, Thunder Bay - Tuesday, Dec. 8

Event Date: 
Tuesday, December 8, 2020 - 11:15am to 11:45am EST
Event Location: 
Thunder Bay, Orillia, Barrie
Event Contact Name: 
Karen Romito
Event Contact E-mail: 

Dr. Gurjit Randhawa
Candidate for Limited Term/Tenure Track Appointment positions, Thunder Bay Campus
Tuesday, December 8, 2020
Zoom Meeting
Meeting ID: 964 3758 9058
Passcode: 093237
Research Presentation - Open to the University Community
Research Title: Machine learning for rapid and accurate genome analysis

In the field of bioinformatics, sequence classification is the scientific practice of identification,
naming, and grouping of organisms based on their similarities and differences. The problem of
sequence classification is of immense importance considering that nearly 86% of existing species
on Earth and 91% of marine species remain uncatalogued. Due to the magnitude of the datasets,
the need exists for a scalable approach that can process large datasets to perform rapid
comparative analysis. I propose ML-DSP and MLDSP-GUI, stand-alone alignment-free software
tools that use Machine Learning and Digital Signal Processing to classify genomic sequences.

MLDSP classified 7, 396 full mitochondrial genomes at various taxonomic levels, from kingdom to
genus, with an average classification accuracy of > 97%. The use of proposed methods was
extended to successfully address a variety of problems such as virus sub-type classification,
bacterial genome analysis, cancer type classification, mitochondrial disease classification, virushost co-evolution, the evolution of extinct and extant hominids, human haplogroup classification, etc.

My recent publication provides proof of principle that the proposed pipeline can classify
newly discovered organisms by correctly classifying the novel COVID-19 virus (SARS-CoV-2). The
proposed approach being alignment-free does not require gene or genome annotations and can
be useful in delivering highly-accurate real-time and fine precision in taxonomic predictions for
future viral and pathogen outbreaks. The technique is being evaluated to assess spontaneous
mutational mechanisms, compromised DNA repair, and environmental mutagen impacts.

Teaching Presentation: Open to the University Community
Tuesday, December 8, 2020
10:40 am - 11:10 am
Zoom Meeting
Meeting ID: 958 5371 0324
Passcode: 244546
Teaching Presentation Title: Feature detection and description in Computer Vision