More than six months af­ter se­cur­ing a nine-fig­ure deal with Sanofi for tar­get dis­cov­ery, a French AI re­search group has land­ed its next Big Phar­ma part­ner­ship.

Owkin, un­der the lead­er­ship of CEO Thomas Clozel, an­nounced Wednes­day morn­ing that it’s signed on with Bris­tol My­ers Squibb to op­ti­mize the Big Phar­ma’s clin­i­cal tri­als and im­prove the odds of clin­i­cal suc­cess, ini­tial­ly in car­dio­vas­cu­lar dis­ease. The an­nounce­ment comes af­ter a deal Owkin made with Sanofi last No­vem­ber that build on the pair’s pre­vi­ous dis­cov­ery col­lab­o­ra­tion.

In Wednes­day’s deal, Owkin gets $80 mil­lion in a com­bined up­front and eq­ui­ty pay­ment and po­ten­tial­ly fur­ther pay­ments high­er than $100 mil­lion as clin­i­cal tri­als move in­to lat­er stages. The biotech added in a state­ment that it plans to use the funds to sup­port its da­ta gen­er­a­tion strat­e­gy in mul­ti­ple ther­a­peu­tic ar­eas, pri­mar­i­ly fo­cus­ing on sin­gle-cell omics.

The new col­lab­o­ra­tion dif­fers from the Sanofi deal. In­stead of us­ing its AI plat­form to im­prove tar­get dis­cov­ery ef­forts and seek bet­ter drug com­bi­na­tions, Owkin is pure­ly fo­cused on clin­i­cal tri­al de­sign and op­ti­miza­tion.

Clozel said the com­pa­ny is fo­cus­ing on re­fin­ing the end­points in tri­als, find­ing the right peo­ple and re­duc­ing tri­al “noise.” And as part of its work, there are a few things he wants to keep in mind.

First is find­ing the right sub­groups and defin­ing them with bio­mark­ers. The sec­ond step, ac­cord­ing to the CEO, is to try and in­tro­duce “syn­thet­ic con­trol arms,” which he de­scribed as more ef­fi­cient ways to op­ti­mize a study’s con­trol arm.

“What we do is try to build a more fed­er­at­ed ac­cess to da­ta world­wide, and re­al­ly bring a new way to match pa­tients with mul­ti­ple modal­i­ties, not on­ly in clin­i­cal tri­als, but to im­prove the way you build for the con­trol arms. That’s the sec­ond one,” Clozel said.

The CEO then went on to talk about how im­prov­ing the con­trol arms can help cut costs with­in clin­i­cal tri­als, af­ford­ing the pa­tients in­volved the op­por­tu­ni­ty to take a treat­ment.

Es­sen­tial­ly, fed­er­at­ed learn­ing works by up­load­ing cer­tain pre­dic­tion mod­els to these servers where da­ta are kept — in this case, pa­tient da­ta — and bring­ing the al­go­rithms back to a cen­tral lo­ca­tion, Clozel said. The mod­els are fine-tuned as need­ed, and then Owkin up­loads the new and im­proved pre­dic­tion mod­els back to the servers. This process re­peats over and over, and the da­ta re­main “locked,” stay­ing in one place.

And with this plat­form, it would work to iden­ti­fy bet­ter end­point de­f­i­n­i­tions, op­ti­mize and iden­ti­fy pa­tient sub­groups with dis­tinct bio­mark­ers, and work on es­ti­mat­ing drug can­di­date ef­fi­ca­cy.

“What we’re try­ing to do is re­al­ly bring­ing the plat­form to help with the things that we think can be op­ti­mized,” Clozel told End­points News.

The third thing that Owkin hopes to ac­com­plish is to di­min­ish the ac­tu­al rigid­i­ty of clin­i­cal tri­als, or the “noise,” us­ing a sys­tem called AI-en­abled co­vari­ate ad­just­ments. Clozel said it can al­low for more ef­fi­cient tri­als and im­prove the p-val­ue.

Wednes­day’s deal fol­lows three years of col­lab­o­ra­tion be­tween Owkin and Bris­tol My­ers, ac­cord­ing to Clozel, in­clud­ing a num­ber of projects to iden­ti­fy bio­mark­ers and im­prove clin­i­cal tri­al out­comes us­ing re­al-world da­ta.

Venkat Sethu­ra­man

Bris­tol My­ers SVP of glob­al bio­met­rics and da­ta sci­ences Venkat Sethu­ra­man said in a state­ment that “We look for­ward to col­lab­o­rat­ing with Owkin to ex­tend our ap­pli­ca­tions of AI and ma­chine learn­ing to en­hance our dis­cov­ery and de­vel­op­ment pro­grams with high-qual­i­ty, di­verse da­ta from clin­i­cal and re­al world sources.”

In its ear­li­er pact with Sanofi, the Big Phar­ma gave Owkin $90 mil­lion just for ex­clu­siv­i­ty for the first three years of the deal, plus $180 mil­lion in an eq­ui­ty in­vest­ment and more on the line in mile­stone pay­ments. That deal was fo­cused on four on­col­o­gy tar­gets: NSCLC, triple neg­a­tive breast can­cer, mesothe­lioma and mul­ti­ple myelo­ma, Owkin told End­points at the time.

Another biotech standing on shaky financial legs has fallen victim to the bears.

Merck partner 4D Pharma has reported that a key lender, Oxford Finance, shoved the UK company into administration after calling in a $14 million loan they couldn’t immediately make good on. Trading in their stock was halted with a market cap that had fallen to a mere £30 million.

“Despite the very difficult prevailing market conditions,” 4D reported on Friday, the biotech had been making progress on finding some new financing and turned to Oxford with an alternative late on Thursday and then again Friday morning.

Ireland has been on a roll in 2022, with several large pharma companies announcing multimillion-euro projects. Now AstraZeneca’s rare disease outfit Alexion is looking to get in on the action.

Alexion on Friday announced a €65 million ($68.8 million) investment in new and enhanced capabilities across two sites in the country, including at College Park in the Dublin suburb of Blanchardstown and the Monksland Industrial Park in the central Irish town of Athlone, according to the Industrial Development Agency of Ireland.

Amidst recently broader vaccine manufacturing initiatives from the EU and European companies, the G7 summit in the mountains of Bavaria has brought about some positive news for closing vaccine and medical product manufacturing gaps around the globe.

According to a statement from the White House, the G7 leaders have formally launched the partnership for global infrastructure, PGII. The effort will aim to mobilize hundreds of billions of dollars to deliver infrastructure projects in several sectors including the medical and pharmaceutical manufacturing space.

Some Celgene shareholders aren’t happy with how Bristol Myers Squibb’s takeover went down.

On Friday, a New York federal judge ruled that they have a case against the pharma giant, denying a request to dismiss allegations that it purposely slow-rolled Breyanzi’s approval to avoid paying out $6.4 billion in contingent value rights (CVR).

When Bristol Myers put down $74 billion to scoop up Celgene back in 2019, liso-cel — the CAR-T lymphoma treatment now marketed as Breyanzi — was supposedly one of the centerpieces of the deal. After going back and forth on negotiations for about six months, BMS put $6.4 billion into a CVR agreement that required an FDA approval for Zeposia, Breyanzi and Abecma, each by an established date.