Case study: Open Targets at your fingertips with Raycast

This blog post is part of a series that will explore applications and expansions of the Open Targets informatics ecosystem, particularly the Open Targets Platform and Open Targets Genetics through conversations with our users.

Irene Lopez Santiago is a bioinformatician in the Open Targets core team. With a background in pharmaceutical science, Irene is constantly looking for ways in which technology can aid scientists in their everyday work. "At Open Targets, I find the perfect balance between both things because we benefit from the many new developments in target identification and technology to build an open platform," she says.

Open Targets’ data team are the governors of the data within the Open Targets informatics platforms. With every potential new source of data, the data team oversees its entire cycle within our ecosystem: from investigating how it would fit into our prioritisation algorithm, modelling the data so it follows our schemas, to ingesting it, and once integrated, fixing and improving the data based on user feedback.

If you need to query Open Targets platforms as often as Irene does, it quickly becomes a very repetitive, labour-intensive process. So she thought to take advantage of the capabilities of the Raycast productivity app to create a simple, fast interface to the data. The app allows users to search for any of the entities in the Open Targets Platform and Open Targets Genetics, browse the available datasets, and access our documentation, all within a single interface.

"A lot of our research requires a detailed look at our data, so having a simple way to see what gene an Ensembl ID refers to, for example, is a small improvement that speeds up the process. The functionality that I use the most at the moment is the ability to check the schemas of each of our datasets," Irene explains. "In Open Targets we have a lot of information, so remembering how each piece is encoded is difficult. The extension is handy for this kind of problem."

Was there anything unexpected or particularly difficult about creating this application?

My lack of experience in web development was definitely a challenge. The extension is written in Typescript, so it was very interesting to see how difficult it was to set up the programming environment, or to do something as simple as an API query. I am used to programming in Python, where asynchronous calls and component states are not a thing. In general, it was very fun to see myself as a toddler developer again.

Where do you think the limits of this application are?

Raycast is a very nice framework because you can quickly provide a front end to your application that is intuitive and feature rich at the same time. It is very good if you want to display text or small amounts of information, but at the same time since it is an extension, it is not meant to deal with complex data representations. In other words, we don't want to use Raycast to replicate our informatics platforms, just to make the interface between the user and our products more seamless.

What are your plans to develop this further?

The next feature I plan to add will be the ability to ask questions to our documentation with natural language. As a first step, it would point the user to the relevant section of our documentation. The next step would be for the app to generate an answer based on multiple sources including our documentation, but also our publications, and ideally, topics within our GitHub and the Open Targets Community.

Read more about the app and how Irene developed it: https://blog.irenelopez.dev/the-gateway-to-open-targets-you-didnt-know-you-needed