Announcing MELLODDY, a consortia of pharma, technical, and academic partners coming together to promote drug discovery with AI and blockchain
Here's an alternative link to an article in the Financial Times
rchestration for D
project uses federated learning and a blockchain to allow traditionally competitive businesses to collaboratively train an AI model used for drug discovery. Here's the amazing thing: these companies are collectively training an AI on their highly confidential data but are guaranteed that the privacy of that data won't be compromised.
It is precisely the kind of paradigm shifting use of new privacy preserving technologies that I laid out two weeks ago
and that we can expect to see more of in the future. To learn more about federated learning I'd suggest this blog post
or this high level non-technical comic
10 of the largest pharma companies in the world are participating, and they are joined by 5 industry partners and 2 academic universities. MELLODDY is taking place under the auspices of the Innovative Medicine Initiative
, and is funded to the tune of €18.4 million. As part of the IMI process these projects are socialized among relevant companies to see who is interested in participating, and it's interesting to note the big pharma companies which (presumably) declined to join.
Bringing this back to blockchain, a private blockchain is used to orchestrate the federated learning process and keep a record of what data has been accessed. That ensures that no single party has control of the learning process, and that it is auditable by each participant. It's an interesting use case, and I'd like to better understand what the tradeoffs are between this approach, and say, Google's non-blockchain federated learning for autocorrect.
As far as I know It's not clear the specific data MELLODDY is targeting, but I'll be keenly watching the outcomes of this project.Austin Blockchain Collective launches a healthcare working group
The Austin Blockchain Collective boasts some 140 blockchain and
crypto companies local to Austin. They just launched a healthcare working group with initial membership from Amchart
and Dell Medical School. This announcement pairs well with John Bass' reflections on Austin
back in May.Walmart Joins Pharmaceutical-Tracking Blockchain Consortium MediLedger
The world's largest retailer and a pharmacy giant has joined the MediLedger project, which is building blockchain solutions for the pharma industry. Previously Walmart's involvement with blockchain had been primarily through their collaboration with IBM on FoodTrust, a track and trace solution for fresh produce built on Hyperledger Fabric.
Details on Walmart's exact participation is sparse but my guess is that Walmart is involved in some way with MediLedger's saleable returns verification
use case and perhaps their DSCSA pilot. Strategically this is important for MediLedger as it breaks Walmart away from their previous tie to Farbic, and officially involves the first pharmacy in their ecosystem. Solve.Care Partners with Leading Pharmaceutical Company Boehringer Ingelheim to Launch Care Network for Patients with Diabetes
The press release is, like most, a bit hand wavy on the details but it sounds like this collaboration will be around education and marketing. It's not clear to me how / if blockchain is being used. For HHS, blockchain means faster ID management and safer mangoes
This brief article features two leaders in senior positions in the Federal Government who have hands on experience with leading blockchain projects and are stalwart proponents for the technology. Frank Yiannis, the FDA’s deputy commissioner for food policy and response, was previously at Walmart where he heavily involved in Walmart's activity in Foodtrust. Jose Arrieta, the new CIO of the HSS, previously led the successful HHS Accelerate program. Both share that they are actively pursuing blockchain opportunities, with Jose Arrieta providing some details about the cross agency projects he has on the horizon. Molecule wants to change how new drugs are funded and developed. Here's an example of how.
Molecule is betting on new token mechanisms and incentive structures to help open up science and disrupt how drugs are developed. The above article gets into some of the details and numbers of how this is operationalized. I'm a huge fan of the open source movement and experiments with cutting edge technology, so I'm really excited by what they are doing, and I hope this experiment is successful.
Like many things in crypto, their solution has a lot of complexity and technical details necessary to make it work. It'll take some amazing UX to abstract away all of that math and make an intuitive user experience.
Here's a follow up article
, also by Molecule, explaining why securitizing pharma IP makes sense, and where they add value.
Podcast: Allscripts hosts Hashed Health CEO John Bass in their "Blockchain for Healthcare" episode