Just-DNA-Seq is an open-source personalized genomics platform designed to empower individuals with control over their genetic information. Building on our successful implementation of previous GitCoin-funded initiatives, we propose to expand and enhance our project in key areas to further advance personalized longevity genetics.

Achievements to Date:

• Source Code and Tools: Developed 23 source code repositories (GitHub).
• Oak Var Modules: Published 11 Oak Var modules focusing on longevity and disease risk.
• Genetics of Aging and Disease Report: Open-source reports detailing health risks and personalized longevity recommendations (example report).
• VO2 max trainability module.
• Genetics Genie GPT (Chat with Genetics Genie) to help people answer genetic questions.
• User Education and Community Engagement: Created a YouTube channel (YouTube) with workshops and lectures about Genetics and LLM.
• Research and Publications: Published a scientific article preprint (arXiv).
• Standalone AI assistant with genetics knowledge supporting Llama3.1 and other open-source models (most recent achievement).

When you sequence your DNA or that of your loved ones, you're left wondering how your genetic code affects your health, intelligence, and other traits. The problem lies in three key areas:

Transparency and trust

Commercial companies often use proprietary algorithms and databases, making it impossible to understand how they arrive at their conclusions. It's like trying to decipher a code without the key. And, to make matters worse, different companies may give you conflicting results.

The solution for the lack of transparency and trust is already deployed. Our open-source platform (Just-DNA-Seq) is available for anyone to use, and you can run it on your laptop instead of our server. You can also check our code and trace how each prediction was made. All the code and databases are open-source and available on our GitHub. You can even put your own filters and explore your genome in detail instead of just reading our reports. However, our output is still somewhat technical for non-professionals and we need to improve it and add more tutorials and documentation.

Explainability

Traditional tools struggle to make sense of the vast amounts of genetic data. It's like trying to find a needle in a haystack. New machine learning models can handle this complexity. However, these models often come as 'black boxes'—mighty yet impenetrable, offering little to no insight into which biological processes are involved.

The solution for the lack of explainability is already in progress. We've started implementing GenNet, an interpretable neural network model designed for genomic data, into our personalized genomic service. We've also gained access to the UK Biobank, which we plan to use to train new models. To continue our efforts, we need your support.

Requirements for personal expertise

While you may know much about yourself and your relatives, you're not a geneticist. Interpreting the results of your genetic testing requires specialized knowledge. Even when the results are transparent, it's like trying to read a foreign language without a translator.

The solution to the lack of personal expertise is underway. We've already developed Genetics Genie. It is available as standalone chat with Llama3.1 and other open-source models (Chat with Genetics Genie) and also as custom GPT (Genetics Genie on OpenAI). However, we want to increase the number of genetics databases it knows and improve its explanations. With your support, we can continue to develop AI assistants that provide clear and concise explanations, making genetic knowledge accessible to everyone.

By supporting our open-source project, you'll be helping to create a more transparent, explainable, and accessible way to understand the secrets of your genome. Join us in democratizing personalized genomics and empowering individuals to take control of their health.