'Smoother': Analyzing 3D genome architecture data on-the-fly, powered by bokeh

Hey everyone,

Using bokeh, we have built an app to analyze Hi-C data. It is called “Smoother”.

Hi-C data is basically a map of how our genomes fold up in the cell. The data is displayed as a heatmap, where each bin of the heatmap expresses the probability that two loci (i.e. positions) of the genome are close to each other in 3D space. Such data is important in biology because it helps us understand how genes are regulated and how cells work.

Smoother is a tool that lets you analyze Hi-C data on-the-fly. It’s as smooth as it sounds! It uses a bokeh server for an interactive interface: Just click some buttons or adjust some sliders to filter and normalize your Hi-C data interactively. No more waiting around for lengthy computations. Under the hood, we use a prefix-sums index to process hundreds of gigabytes worth of data within seconds.

Here is a gif of Smoother in action:
smoother in action

You can check it out on GitHub right here: Smoother GitHub, or read about the algorithmics that make this possible here: Smoother Publication.