MIT researcher creates Dash machine learning tool, study published in Nature

The Dash application wraps a beautiful user interface around the inDelphi machine learning algorithm, a tool for scientists editing genes with CRISPR.

"Dash has been a great support for my scientific work."

Dash's flexible Python framework makes it a powerful tool for scientific research and discovery. A recent study published in Nature, the international scientific journal, utilizes a machine learning tool built with Dash to help scientists predict the outcomes of certain genome editing methods.

The machine learning algorithm, called inDelphi, is published as a Dash application built by Max W. Shen, a computational biology PhD candidate at MIT.

Dash brings data to life. Pictured here, the inDelphi algorithm predictions of gene editing outcomes. (Image source: Max W. Shen).

Shen, with Mandana Arbab, PhD, of Harvard University, and eight other researchers, trained the inDelphi algorithm to predict the results of a specific mechanism of DNA repair called end-joining, which is known to create unpredictable and undesirable insertions and deletions during the genome editing process. These insertions and deletions can result in cell death. inDelphi makes this process more predictable, controllable and useful for scientists doing genomic editing.

Dash applications are built with reactive components, like the dropdown filters above, giving end users real-time control over their data views. (Image source: Max W. Shen).

"Dash will remain my top choice for building web apps."

Shen had worked before with computational biology tools that weren't user-friendly, and he didn't want to feel like he was "fighting the code" just to do his research. "I wanted to create a web app that is a joy to use. As a Python user, Dash was the obvious choice," said Shen.

Creating the initial, functioning inDelphi Dash app took just a few days, and Shen used the next several weeks to create a polished, elegant end-user experience.

"Not only has Dash been a great support for my scientific work, Dash has been artistically fulfilling to use," said Shen. "It was amazing to build a functioning web app in just a handful of days, especially one that was slick and beautiful straight out of the box."

About Dash

Dash is a Python framework that empowers data scientists, researchers, and analysts to build web-based analytics apps without JavaScript. So whether you’re doing computational biology, genomics, or predictive business analytics — wrap beautiful, intuitive and interactive user interface around your algorithms and Dash into your data future.

Interested in Dash?