Inference of CRISPR Edits (ICE)

The Problem: CRISPR allows rapid, highly efficient genome editing - but Next Generation Sequencing is typically slow and expensive. Sanger sequencing is must cheaper and faster, but traditionally does not yield very detailed results.

The Solution: Utilizing a lasso regression machine learning model, ICE is able to extract much more accurate and detailed data from Sanger data than was previously thought possible.

Fast, Accurate, and Free to the Public
Available as a publicly-accessible. cloud-hosted web-tool, ICE allows users to upload any number of files, and can complete its analysis in seconds. Clear and easily-interpretable visualizations offer users ready-made figures suitable for use in papers and publications. The tool was developed to support both mobile devices and desktop/browser interfaces.

By making ICE free to the public, the company encourages not only the adoption of CRISPR technologies by academic labs and smaller research organizations, but provides a reliable point of comparison between the company’s own products and competing methods.

My Role in the Project:
Product Manager, Software: I led software, data, and commercial teams from Development through Release.

Although I had never formally studied the field, I gained a rapid understanding of CRISPR editing and bioinformatics in the space of three months, even speaking at the Keystone Symposium and Harvard University's Department of Molecular and Cell Biology as the product representative for ICE.

Metrics of Success:
- First public release in April 2018 (four months after joining the company as PDM)

- Over 100,000 separate analyses within the first year, leading to approximately 1,400 new orders for the company’s CRISPR products
Industry-leading tool, cited in over 100 publications by the end of 2019