Insilica is thrilled to announce that we have been awarded a grant from the NSF for as part of the Prototype Open Knowledge Graph project. Our data dependency system, BioBricks.ai will be extended to form an open knowledge graph, BioBricks-OKG, for chemical safety. BioBricks-OKG is one of 20 teams building knowledge graphs for this NSF projects, the teams will harmonize data from diverse domains like criminal justice, education, supply and demand, and many more (award list). The goal of Proto-OKN is to provide a unifying fabric for all of these knowledge graphs. If successful, Proto-OKN will provide a common language for data scientists and application developers across a massive domain.
What is BioBricks-OKG?
BioBricks-OKG is designed to revolutionize the way we access and understand chemical health and safety data. It aims to modify our existing open-source system, BioBricks-AI, turning it into a knowledge graph that offers quick access to normalized chemical safety relevant data. This enhancement will make vital safety more accessible and interoperable. It will provide developers a common language on which to build the an ecosystem of applications for chemical saftey and empower AI tools to help develop new and efficient methods for chemical testing and regulations.
Currently, BioBricks-AI provides over 60 public health and cheminformatics databases. Our goal with BioBricks-OKG is to consolidate this data into a single, large, knowledge graph. This graph willassociate chemicals, health risks, testing methods, and various other essential safety entities.
Our partner, NICEATM (National Toxicology Program’s Interagency Center for the Evaluation of Alternative Toxicological Methods), will be helping guide this technology to address needs in the toxicology space. There is a growing need for a data ecosystem to stitch together new approach methodology (NAM) technologies spanning in vitro, in silica and in chemico methods. It is our hope that BioBricks-OKG will provide a foundation for this growing ecosystem.
Why should I care?
You live in a chemical world, chemical safety impacts the air we breath, the water we drink, drugs, food, cosmetics, once you start thinking about it, chemicals and chemical safety is really of critical importance to every part of our lives. It is also a domain with a massive amount of data. However, much of this data remains trapped in isolated repositories, making it a challenge for developers to access. BioBricks-AI has already made strides by offering a unified way to distribute data, simplifying the process of importing databases. But with BioBricks-OKG, we’re aiming to tackle an even bigger challenge: harmonizing the immense volume of available data into a singular knowledge graph. This graph will enable users to quickly find, understand, and apply chemical safety data, ensuring safer practices and better-informed decisions in the cheminformatics space.
Right now, chemical safety data is extraordinarily heterogenous, people describe chemical hazards in different ways, testing methods are sometimes ambiguous, even chemical formulations can be inconsistent. BioBricks-OKG will harmonize this data, making it easier to understand and apply. This will make it easier for developers to build applications and reduce redundancy in the application and data space.
What about LLMs?
This is a great time to be building large language models! Recently a few groups have demonstrated that large language models love knowledge graphs. You can superpower an LLM by allowing it to query a knowledge graph to complete prompts. It feels like LLMs work with every technology, but we think knowledge graphs are special, and the marriage of these two technologies may prove to be a game changer.
A Broader Impact
The potential of BioBricks-OKG isn’t limited to just chemical safety. By partnering with organizations like NICEATM, our knowledge graph can be applied to address various challenges in the chemical domain. Additionally, the graph’s adaptable nature means it can be extended to other areas of health informatics, making diverse data sources more accessible and beneficial.
Presently, BioBricks-AI provides a big data platform for the Ontox Project. With the introduction of BioBricks-OKG, we believe many more will benefit from a unified platform that provides an expansive, harmonized database.
To dive deeper into the specifics of the NSF’s Prototype Open Knowledge Graph project and to get more details on our award, click the links below: