Due to the increasing reliance of scientific research on complex computational pipelines, foundational software in the life sciences has outgrown individual labs and institutes in terms of its scale and impact. This has motivated us to start something more than a code repository for an exciting new project – a new consortium to organize and support core (mostly) Python tools for single cell life science research. We are excited to announce scverse.
What tools are part of scverse?
Our primary experience lies in the field of computational biology with an emphasis on single-cell omics technologies. This experience has defined our initial scope and focus in the very beginning of our journey. Hence, we start with two key data structures for single-cell data in the Python ecosystem, AnnData for uni-modal data and MuData for multi-modal data, together with Scanpy for general purpose single-cell analysis, muon for multimodal single-cell analysis, scvi-tools for deep probabilistic analysis of single-cell data, scirpy for T-cell receptor analysis, and squidpy for spatial omics analysis. We are looking forward to expanding with more major libraries and frameworks which fit our vision soon.
What are we doing?
Our first priority is to build a strong and lasting community around the core tools powering single cell data analysis today. The knowledge to build and maintain these tools should be shared among many to avoid single points of failures. Therefore, we are creating open community channels, creating shared infrastructure to ease package development, and building tooling to make creation of scverse based packages easier. Furthermore, to bring the community together we are planning workshops and hackathons which will be open to the general community. As we grow, we are hoping to recruit more members into governance roles as well as find even more ways to support and connect our community of developers and users.
There are already a few visible outcomes of our joint work including improved mudata performance and mudata documentation. Additionally, scvi-tools is now using MuData for the representation of multimodal data. Scanpy has gained new workshop notebooks, and, moreover, we have started a collection of learning resources for scverse libraries. We’ve also prepared a template for developers to help them to jump-start their new libraries.
Who is a part of scverse?
scverse is explicitly an effort beyond just a single research group. As a community effort we strive for interoperability, shared growth, and democratized governance. In our initial form, we have divided responsibilities into several roles, including:
Isaac Virshup, Danila Bredikhin and Lukas Heumos form the steering committee and also belong to the core developers together with Adam Gayoso, Giovanni Palla, Gregor Sturm, and Ilia Kats. The core developers have made significant contributions to the ecosystem and are jointly developing and maintaining the major libraries.
How can I join and contribute?
We’re particularly excited about the community coming together for this effort with so many brilliant people involved who design tools, write code, manage and analyse data and visualise results. Please ask usage questions on our Discourse forum, developer questions on our Zulip chat and make sure to follow us on Twitter to keep track of scverse updates.
scverse is very much open to contributions in all areas by everyone. We strongly believe in diversity of all kinds being a requirement and a strong benefit for a healthy community. Hence, we embrace people from all backgrounds and experiences, who challenge each other’s assumptions with fresh perspectives. We’re committed to fair treatment and access for all members of the scverse community which is reflected in our code of conduct. A sense of belonging is important to us, so we strive to actively foster a community where everyone is and feels welcomed, respected, supported, and valued. We want to especially encourage people from underrepresented backgrounds to contribute and to reach out.