Next-gen an­ti­bod­ies have be­come some­what of a Holy Grail in the drug de­vel­op­ment world, but the process of ac­tu­al­ly cre­at­ing those an­ti­bod­ies has proven to be dif­fi­cult and time con­sum­ing. A brand-new Cal­i­for­nia out­fit is look­ing to solve that time crunch by bring­ing AI and ma­chine learn­ing in­to the lab.

BigHat Bio­sciences de­buted Wednes­day with a $19 mil­lion Se­ries A round to help scale up its “wet lab” AI and ma­chine learn­ing plat­form to speed de­vel­op­ment of next-gen an­ti­bod­ies.

The San Car­los, CA team led by Google.ai and Broad in­sti­tute vet­er­an Mark De­Pris­to and chief sci­en­tist Pey­ton Green­side, a Stan­ford grad with Broad blood­lines, are aim­ing to bring AI and ma­chine learn­ing in­to the ex­per­i­men­tal process in an at­tempt to dra­mat­i­cal­ly in­crease the speed of an­ti­body dis­cov­ery.

“The re­al goal of our ex­per­i­men­tal plat­form is to en­able us to go from a pre­dic­tion to an in sil­i­co de­sign through to a syn­the­sized, pu­ri­fied and char­ac­ter­ized an­ti­body in a mat­ter of days,” Green­side told End­points News. “What that means is there’s this core work cell that al­lows us to go from DNA to pro­tein very quick­ly and have a very rapid char­ac­ter­i­za­tion of that pro­tein.”

What that looks like in prac­tice is turn­ing out hun­dreds of an­ti­bod­ies in very short or­der — an out­put that could play well in terms of fu­ture out­li­cens­ing deals with bio­phar­ma play­ers. But in the short term, BigHat, which takes its rather nerdy moniker from the “hat” func­tion in sta­tis­tics de­not­ing an es­ti­mat­ed val­ue, will work on de­vel­op­ing and scal­ing its plat­form to give it even more pre­dic­tive and com­pu­ta­tion­al pow­er to pro­duce more an­ti­bod­ies, De­Pris­to said.

The biotech will al­so look to turn drug­mak­er, shep­herd­ing ther­a­peu­tic pro­grams through the clin­ic in the far fu­ture, De­Pris­to said. In terms of what that could look like, De­Pris­to said BigHat was look­ing more at de­sign chal­lenges for an­ti­bod­ies rather than spe­cif­ic ther­a­peu­tic ar­eas, but he did specif­i­cal­ly call out au­toim­mune dis­eases, on­col­o­gy and an­ti-in­fec­tives.

“Be­cause our plat­form can op­ti­mize for pret­ty much any mea­sur­able prop­er­ty of a mol­e­cule — so not just affin­i­ty — we can op­ti­mize for bi­o­log­i­cal func­tion,” De­Pris­to said. “We’re look­ing for pro­grams where the bi­o­log­i­cal ef­fect is not just max­i­mized by high affin­i­ty. The goal is not to per­ma­nent­ly stick the cell to a mol­e­cule but to re­al­ly ac­ti­vate the cell and get its bi­o­log­i­cal func­tion mov­ing.”

With $19 mil­lion in hand from the Se­ries A round led by VC firm An­dreessen Horowitz and joined by 8VC, AME Cloud Ven­tures and In­no­va­tion En­deav­ors, BigHat will look to grow its team of nine em­ploy­ees, adding team mem­bers in the bi­ol­o­gy and com­pu­ta­tion­al spaces. The ex­pand­ed team will add pow­er to BigHat’s plat­form “10 to 100x,” De­Pris­to said.

The com­pa­ny comes with a star-stud­ded sci­en­tif­ic ad­vi­so­ry team, in­clud­ing No­bel lau­re­ate Bri­an Ko­bil­ka, AI and deep learn­ing ex­perts in Jake Gard­ner and An­drew Gor­don Wil­son, and a slew of ex­perts in life sci­ences, en­tre­pre­neur­ship and drug dis­cov­ery. That ad­vi­so­ry team was specif­i­cal­ly con­struct­ed to help BigHat not on­ly scale its plat­form but pre­pare for the tough trans­la­tion­al work that would be re­quired to take its an­ti­bod­ies in­to the clin­ic, De­Pris­to said.

Merck took a big gamble when it opted to jump into the Covid-19 vaccine race late, and made an equally momentous decision to back out in late January. Now, looking to chip in on the effort, Merck reportedly agreed to team up with one of the companies that has already crossed the finish line.

President Joe Biden on Tuesday is expected to announce a partnership between drugmakers Merck and Johnson & Johnson to jointly produce J&J’s recombinant protein Covid-19 vaccine that received the FDA’s emergency use authorization Saturday, the Washington Post reported.

FibroGen will likely have to delay its US rollout for roxadustat once again.

In an unexpected move, the FDA is convening its Cardiovascular and Renal Drugs Advisory Committee to review the NDA in an advisory committee meeting. The date is yet to be confirmed.

Just a few weeks ago, SVB Leerink analyst Geoffrey Porges predicted that the roxa approval could come ahead of the PDUFA date on March 20 — effusive despite already being let down once by the FDA’s extension of its review back in December. AstraZeneca, which is partnered with FibroGen on the chronic kidney disease-related anemia drug, disclosed regulators had requested further clarifying analyses of clinical data.

CRLs. 483s. CBER, CDER and RWE. For biopharma professionals, these acronyms command attention because of the fundamental role FDA plays in drug development. Now Endpoints is doubling down on regulatory coverage, and launching a weekly report focusing on developments out of White Oak, with analysis and insight into what it all means.

Coverage will be led by our new senior editor, Zachary Brennan. He joins Endpoints from POLITICO, where he covered pharma. Prior to that he was the managing editor for Regulatory Focus, a news publication from the Regulatory Affairs Professionals Society.