Menu Close

REU Student Project Abstract

Omkar Haridas Headshot
Omkar Haridas
Faculty Mentor(s): Dr. Matthew Hudson
Algorithm Acceleration Related to Crop Systems
The project I'm working on revolves around a software called CROPSR, which performs CRISPR/Cas9 genome-wide guide RNA design for a given crop genome input, more specifically a .fasta and .gff file. This analysis is recorded in a .csv format spreadsheet for direct viewing, containing the start and end positions of the sequences in the given chromosome, the position of cut sites, on site score for the sequence, and any other features that might be present. The project focuses on software optimization; my goal is to make the software run faster or with less resources to achieve the same result, preferably to achieve this through a reduction in time, as the current version of the software can take days to run. As such, there is a need to reduce this to a scale of hours through optimization. As part of the program, I have been converting the CROPSR code from Python, the language in which it was originally written, into C++. This allows for further scalability of the software and is also meant to be an exercise in optimizing the code, expanding the reach of the devices that it can run on. By extension, another goal of the project is to achieve further optimization through GPU acceleration, which requires the software to be rewritten in C++ first. I have been working under Prof. Matt Hudson for this project, and have learnt a lot through his experience relating to these fields.Throughout the REU program, I have attended various professional development seminars which have given me insight into various important topics that have both directly and indirectly shaped my stance when it comes to this project. The importance of teamwork, entrepreneurship, art of scientific writing, and more have all been imbibed into me as I have applied them to my work. Through this program, I have been in consultation with various students and faculty from Cornell and have discussed various ways that parts of the above task could go about being performed. The nature of this project being collaborative, Cornell is also seeking to achieve similar optimizations, and my C++ rendition of the CROPSR software will play a part in this. Furthermore, through being in contact with other students from the REU program, I have been exposed to a broader range of students than I otherwise would have met in the UIUC CS and CPSC programs, allowing me to assimilate more opinions from a multitude of sources.