Community developers integrate tools into KBase so that reproducible science analysis is freely available
Overview
Five Department of Energy – Office of Science – Genome Science Program’s Science Focus Areas (SFAs, more information here) collaborated with KBase over a 2-year period with the goals of 1) training their developers to use the KBase Software Development Kit (SDK), and 2) integrating their tools and data into the KBase platform to support their science objectives. Overall, the collaborative effort was viewed as highly successful by both the lead researchers and their developer(s). Every SFA project was able to achieve their target goals for training and functionality, while also gaining an appreciation for the role KBase plays/could play in their future science plans.
Impact on the science
In each collaboration, the SFAs were able to provide a portion of developer time to integrate analysis tools into KBase to support the SFA’s science (Table 1). Across the board, this strategy seemed to be well-received, even though it can be challenging to focus on technology-related goals alongside the often complex and ambitious SFA science goals. On average, the SFA developers indicated that they spent ~20% of their overall time on their KBase Apps, acknowledging that their time was sometimes sporadic and other times more intensive. After two years, several SFAs were actively using their new functionality inside KBase, with one acknowledging that the integration has greatly improved their research efficiency. All SFAs recognized the value in enabling the broader research community to also benefit from the additional functionality in KBase.
Table 1: SFAs interviewed that incorporated tools and data into KBase in support of their science goals.
Project Name | Lead | Developer(s) | Collaboration details |
---|---|---|---|
Bacterial:Fungal Interactions | Patrick Chain (LANL) | Mark Flynn, Chien-Chi Lo | www.kbase.us/chain-sfa |
ENIGMA | Eric Alm (MIT) | Anni Zhang | www.kbase.us/adams-sfa |
Microbial Systems Biology for Bioenergy | Rhona Stuart (LLNL) | Patrik D’haeseleer, Jeffrey Kimbrell | www.kbase.us/stuart-sfa |
Systems Biology of Soil Microbiomes | Jennifer Pett-Ridge (LLNL) | Ben Bolduc, Rohan Sachdeva | www.kbase.us/pett-ridge-sfa |
Biomass Deconstruction | Julie Mitchell (ORNL) | Ada Sedova | www.kbase.us/mitchell-sfa |
Overall, the SFAs reported significant progress towards their goals (details in next section). And many of those resources have been shared with the community (Figure 1. SFA projects are highlighted in red). A few example science goals achieved during these collaborations include:
- KBase phage analysis pipeline, with tools run by the community > 300 times. Community webinar – November 2020
- Improved metabolic modeling through combined annotation methods. Community webinar – April 2021
- Exploration of taxonomic assignments using diverse tools. Community webinar – April 2021
Many of the SFAs reported an added bonus to these efforts – increased knowledge of KBase’s resources and community investment, expanding their own possible use of the system. Becoming a KBase Community Developer Community developers get started by exploring the KBase Software Development Kit (SDK; available at www.kbase.us/develop). Each use case may be unique, as developers could have a range of expertise and training, and the integration could use only existing data types (recommended) or need new data types, which requires KBase design support. Either way, we recommend integrating tools that are fully developed and functioning off-system. In addition to the SDK, KBase hosts a KBase User Slack group, with a peer-support network of community developers and KBsae staff to troubleshoot and answer questions. Areas for improvementThe exit interviews also provided valuable feedback from our community developers regarding what was working, and not working, with our SDK training. A few recommendations that we will incorporate into our next SDK refresh include:
Documentation
- Better explanations of the KBase data model and architecture
- Documentation and examples at different levels of integration (novice to expert)
Training
- Provide more regular training and co-working sessions, at least one of each yearly
- Develop additional templates that demonstrate a range of functionality
Interview highlights
In summary, we wanted to leave you with a few interview highlights from our community interviews. Overall, the teams appreciated the opportunity to work with KBase, and we really appreciated their following sentiments:
- “[Respects and admires] KBase for the problem they are attempting to address.”
- “KBase has been a great tool that has allowed the project to develop new models much faster.”
- “KBase makes it easier for the non-bioinfo user, especially to find apps and tools.”
- “The powerful aspects of KBase are that it is very interactive, including pipeline capabilities; creating a community of methods that are connected.”
We look forward to getting to know and help other community developers that want to make tools and analyses available in KBase. Please contact us at engage@kbase.us to get started!