Future Literacy

Applying computational advances to improve drug discovery is one of the most promising areas in healthcare. Data-informed decisions can guide developers toward auspicious drugs by augmenting human acquired wisdom, and away from costly detours by helping to keep biases in check.

OS Fund has been watching these developments very carefully, because such breakthroughs have the potential to help hundreds of millions or even billions of people in great need. This opportunity has attracted numerous companies to try their hand in tackling this complex challenge. After diligencing many efforts, we’ve found a few teams that will break away from the pack.

Leveraging their AI-driven drug discovery platform, twoXAR can rapidly move from input data to outputs that can be tested in vivo while reducing the time and cost associated with drug development.

Here’s why we believe twoXAR’s platform will lead to a portfolio of drug programs unprecedented in scale and scope:

Above all, the proof is in the results. twoXAR’s computational platform has already produced real results in animal studies, not just in one or two cases, but with 30% of their drug candidates showing efficacy across four different diseases. By comparison, traditional discovery methods yield a ~2% success rate at this stage of drug development, and many computational engines talk big, only to fail in the lab. The strength of twoXAR’s validation studies has led to multiple collaborations, including one with pharmaceutical partner Santen to discover new treatments for glaucoma. twoXAR’s team has deep and proven industry expertise not only in computational discovery, but in pharmaceutical development. This knowledge allows twoXAR to be a responsive collaborator supporting preclinical studies and to maximize value for industry partners. In parallel, twoXAR is building its own pipeline focused on diseases with clear paths to the clinic including diseases within dermatology, immunology, oncology and ophthalmology.

Their AI-driven approach allows twoXAR to build this portfolio at a massive scale, with the capacity to identify promising drug candidates for hundreds to thousands of diseases. Because twoXAR’s computational methods are scalable and efficient, they can plug and play disease-specific data to new disease models at speeds and costs previously impossible with conventional technologies. twoXAR has already tackled hepatocellular carcinoma (liver cancer), rheumatoid arthritis, type 2 diabetes, and many more diseases. This is just the start — as the team continues to develop additional statistically-independent methods and incorporate new data the confidence of the predictions will continue to increase.

Because twoXAR’s computational methods are scalable and efficient, they can plug and play disease-specific data to new disease models at speeds and costs previously impossible with conventional technologies.

This scalability and efficiency of twoXAR’s platform allows them to extract high-value drugs from existing compound libraries before other companies in the space. twoXAR can screen compound libraries to identify drug candidates much more efficiently than existing high-throughput screening methods, resulting in lead candidates in a matter of weeks vs. the standard 4–6 years. And twoXAR’s approach has advantages, not only over conventional high-throughput screening, but over other computational engines:

twoXAR is less likely to produce false-positive drug candidates because it uses multiple computational methods to corroborate results. Predicting complex biological interactions is challenging and false positives are a known problem with single computational methods. Other companies in this space tend to use more narrowly focused methods, which increases the chance of incorrect predictions. The degree of confidence is much higher with twoXAR’s predictions because they use multiple unbiased, statistically-independent, and optimized methods — applying multiple lenses to the same data for better results.

Because of its multi data set analysis twoXAR is more likely to identify promising drugs others might miss. twoXAR looks for patterns across diverse and independent data sources of biological, chemical, and clinical data to identify unanticipated drug-disease associations. Most other AI-driven drug discovery companies use fewer data points sources, narrowing their perspective and making them more likely to miss drugs that could work.

Finally, twoXAR increases the chance of success by focusing on the end goal: identifying drugs with the greatest likelihood of treating a disease. Unlike other companies in the space who optimize individual steps of the drug discovery process (leaving the possibility of failure during other steps of the process), twoXAR reduces those failure points by using real-world data to directly predict the likelihood of a drug being efficacious in humans.

OS Fund is proud to invest in twoXAR’s Series A round alongside other investors including Softbank Ventures and Andreessen Horowitz. This funding round will accelerate the discovery and development of desperately needed life-saving treatments. Stay tuned for the next breakthrough from twoXAR!