Over a year ago I embarked on a journey with Samuel Burke to try and solve what we believed was the biggest problem in biomedical science, how to clean up all the bad science we saw permeating the scholarly literature. That journey soon became a company we co-founded called VeracityBio and we have been on a committed quest ever since to build the tools that would be needed to address those issues.
One aspect of this problem we have struggled with has been finding other people who recognized the problem and had the skills/knowledge needed to help us do something about it. Though we did manage to find some incredibly resourceful individuals along the way, such as Matt Carland, Ninaad Kaala and Eric McLachlan, who have helped us tremendously in getting us to where we are, we have yet to find anyone with the requisite resources and drive needed to really push this project forward.
Then, at the end of last week we came across a talk given by Liran Belenzon at Google Cloud. Liran is the CEO of BenchSci, a company we are well aware of as they are one of the few companies in this space we thought might be working on how to solve the lack of veracity in biomedical science. As best we can tell, they are further along than any other company in trying to root out this problem.
BenchSci also has an incredible story to tell. In nine years they went from a few graduate students with an idea to attracting some of the biggest companies and brightest minds on the planet who together are now working on one of the most difficult problems in biomedicine.
We were buoyed to see that we are among some brilliant thinkers who have recognized and are trying to do something about this issue. It is one of the clearest signs we have gotten so far that what we are working on is a truly pernicious and pervasive problem and solving it might just be the game changer we conceived it might be.
Also, thanks to the tireless work of Ninaad, Matt and Eric, we have made considerable progress towards that end. VeracityBio has developed technology to automate and broaden the assessment of medical research reliability, specifically for data linking potential drug targets with diseases. This powers platforms designed to systematically screen therapeutic targets, evaluate risks and ensure the best possible fit between targets and diseases.
The development of new therapeutics faces significant challenges: published research can be inaccurate, and pre-clinical findings rarely translate. VeracityBio’s approach acknowledges the human element in scientific research, understanding the impact of market incentives on researcher behavior. We are translating the patterns that emerge from these incentives into algorithms – “Veracity Modules” – designed to assess the generalizability, robustness, effectiveness, trustworthiness, and transparency of scientific research.
We want to thank Liran and the whole team at BenchSci for devoting themselves to this problem and showing us that there is a path forward in this incredibly convoluted space we both inhabit.