Two of the scientists who won this year’s Nobel Prize for cracking the code of proteins’ intricate structures relied, in part, on a series of computing models that anyone with a computer and the right understanding can access and run.
Their creation, collectively called AlphaFold, quickly and accurately predicts the three-dimensional structure of each of the millions of proteins in the human body, a key to understanding their unique functions. In designing this model, Demis Hassabis and John Jumper relied on the shared work of tens of thousands of scientists and publicly available open source software, underscoring how open science is powering scientific discovery.
Open source software refers to blocks of code that are developed and maintained online for anyone to use, improve, or share as they see fit — all for free. Biomedical scientists depend on these public tools to analyze large, complex data sets, paving the way for the development of new drugs and treatments.
“You would be hard-pressed to find any piece of scientific literature published today that has not been supported in some way by at least a dozen if not more, open source software projects.”
Kate Hertweck, program manager at the Chan Zuckerberg Initiative (CZI)
CZI is a major open source funder that has, more broadly, devoted billions to preventing, curing, or managing all diseases by the end of the century.
Yet despite their near-universal contributions to research and medicine, open source tools are often underfunded and difficult to support through mechanisms that support research, limiting their true potential — an issue CZI has been trying to tackle for the past five years.
A stubborn funding problem
The potential of open source software was felt globally during the Covid-19 pandemic in 2020, when the IQ-TREE software became an indispensable tool in tracing and detecting virus variants. A CZI grant helped IQ-TREE dramatically improve the speed and scalability of its software, alleviating a computational bottleneck that occurred as scientists updated datasets with thousands of new genomes every day.
“IQ-TREE was being used by many researchers to study the evolution of this virus, which helped us make improvements faster,” Dr. Minh Bui, inventor of IQ-TREE, said in an interview with Australian National University.
Yet open source software faces massive funding challenges, especially as projects mature.
Dario Taraborelli, a program officer at CZI, attributes the funding shortfall, in part, to the incentives system within the scientific community. Researchers are rewarded if their papers are highly cited or published in prestigious journals, often leading to career opportunities, research grants, and scientific awards.
Conversely, developing and maintaining software, even if it supports a wide body of research or meets the needs of broad scientific communities, rarely results in similar opportunities. And yet, sustainable funding for essential software tools that power scientific discovery is critical for infrastructure upkeep, community management, documentation, and regular code updates.
“We don’t value software, data, and methods in the same way we value papers, even though those resources empower millions of scientists,” said Taraborelli.
Big players invest in open source software
The federal government recognizes the value of the collaboration these software tools foster. In 2023, the Biden administration launched the Year of Open Science, a focused effort among federal agencies to advance open research practices and invest in critical infrastructure for open science, including open-source software.
On the philanthropic front, CZI established the Essential Open Source Software for Science (EOSS) program in 2019. During the past five years, the program has invested $53 million into the upkeep, growth, and community engagement of critical open source tools as part of CZI’s broader effort to understand the fundamental mechanisms of biology. The EOSS program (which in its most recent funding cycle involved a partnership with the Wellcome Trust and the Kavli Foundation) describes itself as the largest worldwide fund dedicated to the needs of open source communities in biomedicine and the life sciences.
EOSS grants were awarded to 195 mature, widely adopted software projects that need funding to keep evolving, including IQ-TREE and five tools cited in a seminal paper by the Nobel Prize-winning creators of AlphaFold.
Scientists collaborate — funders must, too
Given the behind-the-scenes nature of open source software, its impact is tough to measure using standard metrics of scientific success, such as citations in scientific literature. Researchers rarely cite the software they use in an analysis, and sometimes, there’s no format available for them to do so, according to a CZI report that explores the impact of open source software in the biomedicine community.
To address the challenge of software citation, the report examined the number of formal citations and informal mentions of open source software projects that CZI has helped fund through its EOSS program. It found that its projects were mentioned in 774,050 scientific papers from 2013 to 2022, and 80 percent of those papers were published between 2019 and 2022, corresponding with the timing of five EOSS funding cycles. Mentions for some of the projects funded by CZI grew by orders of magnitude during this period of time.
Software mentions are not a perfect measure of impact, and there’s no way of knowing if the mentions were a direct result of EOSS funding, the report said. But they are a clear indicator that these tools are being adopted rapidly and funding needs to keep pace. Also, the mentions do not account for other markers of success, such as whether CZI funding helped foster diversity among open source software contributors, an important goal of the initiative.
Regarding inclusivity, CZI saw signs of improvement, with 25 percent of grantees reporting that EOSS funding supported activities that made their software and materials more accessible to a wider cross-section of researchers. For instance, 3D Slicer, a visual computing platform, used the support to help translate project documentation into 45 languages. During its first and second years of funding, downloads increased by 72% and 204% in Senegal and Mauritania, respectively. And for the first time, it has developer groups (including female engineers and students) in Africa and Latin America.
Taraborelli said the report confirms that scientific discoveries are a team effort these days, and the funding that goes into them should be, too. A community of funders needs to coalesce around the open software community to help it thrive.
“We’re trying to figure out how to work across funders and institutions in the public and private sectors to build a more sustainable, predictable source of support for some of these projects,” Taraborelli said. “We’re trying to build bridges.”
Read the full CZI report here.
