The Key to Scientific Breakthroughs? Improving Access to Open Source Software
Improving access to open source software tools is a cornerstone of our science work at the Chan Zuckerberg Initiative (CZI). These tools, which are built by scientific communities for scientific communities, can help accelerate the pace of scientific discovery by increasing collaboration and reproducibility.
Across our different focus areas at CZI, we strive to support core open source technologies that will underpin future breakthroughs in biomedical research. To that end, CZI’s Imaging and Open Science programs are supporting 76 new grants from two funding opportunities that are focused on the sustained growth of the scientific open source ecosystem.
Through the second funding cycle of the napari Plugin Foundations grants, 35 developers were each awarded $25,000 to improve the quality and usability of napari plugins on the napari hub. napari is a community-driven platform for browsing, annotating, and analyzing large multi-dimensional images. The napari hub features over 225 plugins that scientists and researchers can use to visualize and analyze datasets more efficiently — deepening their understanding of biological processes within the human body.
CZI’s Essential Open Source Software for Science (EOSS) program launched in 2019 to provide some of the most widely-used scientific open source tools with much-needed support. With the fifth funding cycle of this program, we are awarding $12 million through 40 new grants.
Through a new targeted grant, we’re also supporting the Scientific Python project — an initiative to coordinate development and grow the community of the scientific Python ecosystem, comprising many of the software tools that CZI funds. Over the next two years, this $1.8 million grant will help the Scientific Python project support common infrastructure and practices, accessibility, and interactivity of documentation of core libraries, thereby expanding global participation of scientific communities in using and contributing to Python tools.
Hear from Justin Kiggins, a senior product manager on the Imaging Technology team, and Carly Strasser, a program manager on the Open Science team, about why funding open source software tools is essential to our work at CZI.
How do these funding opportunities support the imaging and open science communities?
Carly: The EOSS RFA’s inception was really about providing funding for projects that may be under-resourced. The projects we’ve funded are typically well-established in the open source community. Take SciPy, for example, an open source library for scientific computing in Python. The vast majority of tools used in the biomedical research community depend directly or indirectly on SciPy — forty percent of EOSS-supported Python software projects list SciPy as a dependency.
But while projects like SciPy are widely-used and established, they may need additional resources for things like community-building, maintenance, creating documentation, or updating the code base. Securing support for these types of improvements can be very difficult for scientific software because these tools are rarely funded. The impetus for EOSS was to recognize the importance of these pieces of software and how funding them could really improve the foundational infrastructure of science.
Justin: Over the last few years, napari has become one of the leading, community-driven platforms for browsing, annotating, and analyzing large multi-dimensional images, which is critical as the size of imaging datasets continue to balloon. The development and enhancement of napari plugins, which are housed on the napari hub, can help scientists and researchers visualize and analyze datasets more efficiently, so they can have a better understanding of human biology.
The Imaging team has been working closely with the napari community, and this is our second round of funding to support plugin developers. Our goal is to help cultivate a strong plugin ecosystem by supporting imaging analysis tools that are already in development. So, we’re funding both the maintenance of existing plugins and converting existing projects into usable tools on the napari platform. Ultimately, we’re hoping to help grow a robust ecosystem of imaging analysis plugins for napari users and biologists at large, which can help lead to new research breakthroughs.
There have been a couple of other exciting moments for the napari community this year as well. napari recently became a NumFOCUS sponsored project, which will enable additional financial support from CZI and other funders; the sponsorship also established an advisory board for napari with institutional funders and partners like CZI. As a result, we’re now able to fund the growing napari plugin ecosystem alongside funding and building the project itself. Our first direct grant to napari will support community-building activities to help ensure the project’s long-term sustainability. Additionally, we’ve also committed up to $100,000 in the year ahead to match grants from other napari funders.
How are these funding commitments driving efforts to cure, prevent, or manage all diseases by the end of this century?
Carly: We expect that future research discoveries and biomedical breakthroughs may be made possible by some of the tools we’re funding. More generally, these projects are essential to the research communities we support. NumPy is a great example — it’s a core package for numerical and scientific computing with Python. With CZI’s support, NumPy developers launched a new website, offered its first translations in Arabic and Portuguese, expanded its active maintainer group by 60 percent, and gained first-time code contributors from Africa, Asia, and South America. IQ-TREE — a fast algorithm to infer phylogenetic trees and another project supported by CZI — developed an open standard and API for phylogenetic models and dramatically improved its speed and scalability, which proved critical for COVID-19 analysis as datasets got updated with thousands of new genomes per day.
Bringing recognition and resources to open source software in the scientific community is really critical for us because it helps other funders and groups value and understand the role of scientific software in the ecosystem.
Justin: There are a lot of image analysis tools on the market that use paid licensing to incorporate analysis methods. But these specific tools are limited because ‘out-of-the-box’ software solutions can’t necessarily solve the unique challenges facing scientists who are working on the frontiers of imaging and biology.
That’s what makes napari so exciting! By focusing our efforts on an open source, free image analysis platform, academic researchers who are building the latest deep learning methods for finding cells in a brain, for example, are able to do their work and immediately share it with the rest of the imaging community.
Carly: I’ll also just add one more thing. Twenty years ago, science could be done without a lot of attention being paid to the software that was being used. Many scientists used black box software. But today, there’s a wider call from the public around transparency and reproducibility in science — which really speaks to the growing importance of open source software tools. I think it’s going to continue to be really critical for the scientific community to show your work, how you got results, and what you need to use to reproduce these results.
Could you share a bit more about how democratizing access to open source tools like napari is so important?
Justin: There’s a lot of passion and innovation happening in the academic community with respect to image analysis, but many researchers face similar challenges in their labs — such as visualizing and exploring large, complex imaging datasets, or finding the right analysis workflow for their specific experiment. By democratizing access to tools like napari, it can help ensure more knowledge-sharing and collaboration between labs, which ultimately helps lower the barriers to imaging analysis.
We’ve funded a few cycles of the EOSS program — why is it important to continue supporting open source software tools?
Carly: Speaking from my own experience, I learned to code on proprietary software and it wasn’t until after graduate school when I understood the value of having access to free, open source software like Python. Having that access was especially critical for my work after I left an institution that could afford licenses. I’ve been working in the open science space now for about 10 years, and it’s really exciting that CZI focuses on open science. We’re one of the few funders with a specific program focused on open science.