SF tech start­up Atom­wise gets $45M for AI-pow­ered drug de­sign soft­ware

A Sil­i­con Val­ley soft­ware com­pa­ny that’s us­ing ar­ti­fi­cial in­tel­li­gence to take the guess work out of struc­ture-based drug de­sign is trot­ting out a siz­able Se­ries A round this morn­ing. The deal in­cludes in­vestors that run the gamut, from tech to bio­phar­ma to agro­chem­i­cals.

The com­pa­ny is called Atom­wise, and it’s haul­ing in $45 mil­lion in a round led by Mon­san­to Growth Ven­tures, tech in­vestors DCVC (Da­ta Col­lec­tive), and B Group Cap­i­tal – the VC fund found­ed by Face­book co-founder Ed­uar­do Saverin.

Abra­ham Heifets

In biotech mon­ey, $45 mil­lion is start­ing to seem stan­dard if not pal­try for a B round (a sign of the very flush times, no doubt). But in soft­ware, this dol­lar amount still makes a splash, Atom­wise’s CEO and co-founder Abra­ham Heifets tells me.

The com­pa­ny is work­ing on ar­ti­fi­cial in­tel­li­gence soft­ware that will help drug de­sign­ers find mol­e­cules that are worth pur­su­ing – a lengthy process that means time and mon­ey to the in­dus­try.

A lot of com­pa­nies are work­ing to find tar­gets on the bi­ol­o­gy side, Heifets said, ask­ing which pro­teins and path­ways play vi­tal roles in dis­eases.

“But once you’ve an­swered those ques­tions, you need a mol­e­cule that can block that pro­tein or in­hib­it that path­way specif­i­cal­ly. All of that de­sign is on the chem­istry side, and that’s what our soft­ware helps with,” Heifets said.

Right now, ul­tra-high-through­put screen­ing ro­bots do the work to sift through po­ten­tial mol­e­cules hunt­ing for the right puz­zle piece. They can get through about 100,000 com­pounds per day. Atom­wise’s soft­ware can run 10-20 mil­lion per day.

“We’ve seen a shift in the world from scarci­ty to abun­dance,” Heifets said. “We now have syn­the­sized on-de­mand li­braries, and you can or­der some­thing like 600 mil­lion mol­e­cules on the web. But you can’t test 600 mil­lion mol­e­cules phys­i­cal­ly. To grap­ple with the in­cred­i­ble suc­cess of the phar­ma man­u­fac­tur­ing com­mu­ni­ty, you need com­pu­ta­tion­al ap­proach­es.”

Al­though found­ed back in 2012, Atom­wise has seen much of its growth in the past two years. The com­pa­ny struck part­ner­ships with four large phar­ma­ceu­ti­cal com­pa­nies, in­clud­ing Mer­ck, mul­ti­ple biotechs, and over 40 ma­jor re­search uni­ver­si­ties.

Heifets sees the com­pa­ny’s tech as a so­lu­tion to the phar­ma­ceu­ti­cal in­dus­try’s ROI prob­lem.

“The in­dus­try can’t keep go­ing with neg­a­tive ROIs much longer,” Heifets said. “We need new tech­niques. And we think – and have the da­ta to back it up – that AI is a so­lu­tion to those prob­lems.”

That’s a state­ment that might raise some eye­brows in the bio­phar­ma com­mu­ni­ty. There’s a bit of a de­bate over whether com­pa­nies like Atom­wise are over­stat­ing the pow­er of AI to save drug de­sign­ers time.

Sci­ence writer Derek Lowe has tak­en on Atom­wise and oth­er AI com­pa­nies in the past (read more here, here, and here).

“The whole com­pound screen­ing step is just an­oth­er ear­ly thing in pre­clin­i­cal space; I’ve nev­er seen a suc­cess­ful project in which it was a rate-lim­it­ing step. But ‘shave a few weeks off some­thing at the very be­gin­ning’ isn’t as com­pelling an of­fer, is it?” Lowe wrote in a blog a few months ago.

Heifets, how­ev­er, is con­vinced. Af­ter send­ing Lowe’s state­ments to Atom­wise, Heifets wrote me this in an email:

“It’s com­mon­ly un­der­stood that the Lead Op­ti­miza­tion step is the most ex­pen­sive ($414 mil­lion ver­sus a Phase III clin­i­cal tri­al at $314 mil­lion, in 2010 dol­lars).””

Al­though phar­ma is the “big­ger eco­nom­ic op­por­tu­ni­ty,” Heifets said the com­pa­ny sees cus­tomers in both bio­phar­ma and agro­chem­i­cal in­dus­tries.

“The com­put­er doesn’t know or care if the car­bon atom is in­side a hu­man cell or a wheat cell,” Heifets said. “The ques­tion looks the same in bio­chem­istry. Plus, adding agro­chem­i­cal com­pounds makes the tech­nol­o­gy more ro­bust, so we find it valu­able.”

The soft­ware was valu­able enough to Mon­san­to, who led the Se­ries A round.

“We chose to in­vest based on the im­pres­sive re­sults we saw from Atom­wise in our own hands,” Kier­sten Stead, part­ner at Mon­san­to Growth Ven­tures, said in a state­ment. “Atom­wise was able to find promis­ing com­pounds against crop pro­tec­tion tar­gets that are im­por­tant ar­eas of fo­cus for agro­chem­i­cal R&D.”

Atom­wise, which em­ploys 17 to­day, will use the new cash to scale, most like­ly dou­bling its staff by the end of the year and push­ing the lim­its on the soft­ware’s ca­pa­bil­i­ties.

Hal Barron, GSK

Break­ing the death spi­ral: Hal Bar­ron talks about trans­form­ing the mori­bund R&D cul­ture at GSK in a crit­i­cal year for the late-stage pipeline

Just ahead of GlaxoSmithKline’s Q2 update on Wednesday, science chief Hal Barron is making the rounds to talk up the pharma giant’s late-stage strategy as the top execs continue to woo back a deeply skeptical investor group while pushing through a whole new R&D culture.

And that’s not easy, Barron is quick to note. He told the Financial Times:

I think that culture, to some extent, is as hard, in fact even harder, than doing the science.

Endpoints News

Basic subscription required

Unlock this story instantly and join 55,200+ biopharma pros reading Endpoints daily — and it's free.

Michel Vounatsos, Getty Images

UP­DAT­ED: Stay tuned: Bio­gen’s num­bers are great — it’s their wor­ri­some fu­ture that leaves an­a­lysts skit­tish

Biogen came out with an upbeat assessment of their Q2 numbers today, discounting the arrival of a key rival for its blockbuster Spinraza franchise. But the top execs remain grimly determined to not say much anything new about the sore points that have dragged down its stock, including the future of its big investment in Alzheimer’s or how it plans to invest the considerable cash that the big biotech continues to reap.

Endpoints News

Basic subscription required

Unlock this story instantly and join 55,200+ biopharma pros reading Endpoints daily — and it's free.

PACT Phar­ma says it's per­fect­ed the tech to se­lect neoanti­gens for per­son­al­ized ther­a­py — now on­to the clin­ic

At PACT Pharma, the lofty goal to unleash a “tsunami” of T cells personalized for each patient has hinged on the ability to correctly identify the neoantigens that form something of a fingerprint for each tumor, and extract the small group of T cells primed to attack the cancer. It still has a long way to go testing a treatment in humans, but the biotech says it has nailed that highly technical piece of the process.

UP­DAT­ED: My­ovan­t's uter­ine fi­broid drug looks com­pet­i­tive in PhI­II — but can they van­quish mighty Ab­b­Vie?

Vivek Ramaswamy’s Myovant $MYOV has closely matched its positive first round of Phase III data for their uterine fibroid drug relugolix, setting up a head-to-head rivalry with pharma giant AbbVie as the little biotech steers to the market with a planned filing in Q4.

Here’s how Myovant plans to prevail over the AbbVie $ABBV empire.

In the study, 71.2% of women receiving once-daily relugolix combination therapy achieved the clinical response they were looking for, compared to only 14.7% in the control arm. The data comfortably reflected the same outcomes in the first Phase III — 73.4% of women receiving once-daily oral relugolix combination therapy achieved the responder criteria compared with 18.9% of women receiving placebo — which will reassure regulators that they are getting the carefully randomized data that qualifies for the FDA’s gold standard for success.

Lit­tle Mar­i­nus sees its shares eclipsed as the Sage ri­val fails to com­pare on PPD in PhII

The executive team at Sage $SAGE have skirted another potential pitfall on its way to racking up a big future for its depression drug Zulresso.

Little Marinus Pharmaceuticals $MRNS had sought to challenge the Sage drug with an IV formulation — followed by an oral version — of ganaxolone for postpartum depression. But researchers say their Phase II study failed to positively differentiate itself from a placebo at 28 days — leaving them to hold up “clinically meaningful” data within the first day of administration compared to the control arm.

Roche cuts loose Tam­i­flu OTC rights, hand­ing Sanofi the keys as the phar­ma gi­ant dou­bles down on Xofluza

Roche set out to make a better flu medicine than Tamiflu as that franchise was headed to a generic showdown. Now they’ll see just how well Xofluza stacks up against the mainstay drug after handing off over-the-counter rights in the US to Sanofi.

Sanofi $SNY says it will now step in to negotiate a deal with the FDA to steer Tamiflu into the OTC market, a role that could well involve new studies to ease passage of the drug out of doctor’s hands and into the consumer end of the market. And the French pharma giant will have first dibs over “selected” OTC markets around the world as they push ahead.

Aca­dia is mak­ing the best of it, but their lat­est PhI­II Nu­plazid study is a bust

Acadia’s late-stage program to widen the commercial prospects for Nuplazid has hit a wall. The biotech reported that their Phase III ENHANCE trial flat failed. And while they $ACAD did their best to cherry pick positive data wherever they can be found, this is a clear setback for the biotech.

With close to 400 patients enrolled, researchers said the drug flunked the primary endpoint as an adjunctive therapy for patients with an inadequate response to antipsychotic therapy. The p-value was an ugly 0.0940 on the Positive and Negative Syndrome Scale, which the company called out as a positive trend.

Their shares slid 12% on the news, good for a $426 million hit on a $3.7 billion market cap at close.

Endpoints News

Basic subscription required

Unlock this story instantly and join 55,200+ biopharma pros reading Endpoints daily — and it's free.

Some Big Phar­mas stepped up their game on da­ta trans­paren­cy — but which flunked the test?

The nonprofit Bioethics International has come out with their latest scorecard on data transparency among the big biopharmas in the industry — flagging a few standouts while spotlighting some laggards who are continuing to underperform.

Now in its third year, the nonprofit created a new set of standards with Yale School of Medicine and Stanford Law School to evaluate the track record on trial registration, results reporting, publication and data-sharing practice.

Busy Gilead crew throws strug­gling biotech a life­line, with some cash up­front and hun­dreds of mil­lions in biobucks for HIV deal

Durect $DRRX got a badly needed shot in the arm Monday morning as Gilead’s busy BD team lined up access to its extended-release platform tech for HIV and hepatitis B.

Gilead, a leader in the HIV sector, is paying a modest $25 million in cash for the right to jump on the platform at Durect, which has been using its technology to come up with an extended-release version of bupivacaine. The FDA rejected that in 2014, but Durect has been working on a comeback.