Dyno sketch­es tens of thou­sands of vi­able next-gen AAV cap­sids with the help of ma­chine learn­ing — pa­per

Since its launch in May 2020, Dyno Ther­a­peu­tics has tout­ed its plat­form’s po­ten­tial to dis­cov­er vi­able vari­a­tions of cur­rent-gen AAVs with no­tice­able lim­i­ta­tions in terms of pa­tient safe­ty and ef­fi­ca­cy. Now, ac­cord­ing to a new study, Dyno has used its neur­al net­work to out­line tens of thou­sands of vari­ants that could add weight to its mis­sion to build a bet­ter cap­sid.

In a study meant to de­ter­mine how many vi­able vari­ants of the AAV2 cap­sid it could de­sign with the aid of ma­chine learn­ing, Dyno sketched out more than 100,000 virus­es that could be used to car­ry gene ther­a­pies, ac­cord­ing to a new pa­per pub­lished in Na­ture Biotech­nol­o­gy.

Us­ing a neur­al net­work to de­sign se­lec­tive­ly mu­tat­ed sites on a 28-amino acid chain, re­searchers at Dyno iden­ti­fied 110,689 vi­able vari­ants of the AAV2 cap­sid — a suc­cess rate of more than half of all vari­a­tions the biotech’s ma­chine learn­ing plat­form came up with as part of the ex­per­i­ment.

Sam Sinai

How did Dyno’s tech ac­com­plish that feat with lim­it­ed in­struc­tion­al da­ta? Ac­cord­ing to co-founder Sam Sinai, Dyno’s team worked smarter and not hard­er with the da­ta sets avail­able, se­lec­tive­ly in­putting — and some­times omit­ting — da­ta to cre­ate a bet­ter pre­dic­tion on less in­for­ma­tion.

“We looked at how dif­fer­ent ways at look­ing at the same da­ta — or even ig­nor­ing da­ta that we had — can help cer­tain ma­chine learn­ing mod­els in their abil­i­ty to mod­el the space that we are try­ing to go in­to,” Sinai told End­points News. “This is a huge ad­van­tage, re­mov­ing the bur­den of ex­pen­sive ex­per­i­ments from the lab­o­ra­to­ry to the com­put­er.”

The re­sult was a rich va­ri­ety of AAV vari­ants that could of­fer the need­ed di­ver­si­ty to tack­le Dyno’s sig­na­ture chal­lenge — work­ing around pa­tients’ nat­ur­al im­mu­ni­ty to spe­cif­ic AAV serotypes due to pri­or ex­po­sure. Paired with oth­er re­search in­to how to de­sign a cap­sid to bet­ter tar­get spe­cif­ic tis­sues, Sinai ar­gued his team is piec­ing to­geth­er a po­ten­tial road map for the fu­ture of AAV-based gene ther­a­pies.

The ex­per­i­ment has a side ben­e­fit, Sinai said, as one of the biggest ex­per­i­ments ever un­der­tak­en to dra­mat­i­cal­ly re­work the shape of a pro­tein. The high rates of suc­cess in find­ing vi­able vari­a­tions is just the cher­ry on top.

“The study it­self is one of the largest de­signs of any pro­teins to date with ma­chine learn­ing in terms of band­width and in terms of how much change we have in­duced in this pro­tein,” he said. “In that sense it’s very ex­cit­ing. When we start­ed this study in 2017, we didn’t know we could change the pro­tein as much as we did. One of the ex­cit­ing re­sults of this study is that we could.”

With the high suc­cess in map­ping AAV2, Sinai said Dyno is al­so fo­cus­ing on ad­di­tion­al serotypes, in­clud­ing AAV9 — the tech be­hind No­var­tis’ Zol­gens­ma. The team’s com­pu­ta­tion­al pow­er should work the same way across all wild serotypes, Sinai said, which could churn out mil­lions of unique vi­able vari­ants down the road.

So far, some big-name play­ers in phar­ma are tak­ing a bet on Dyno’s grow­ing po­ten­tial, with the Roche/Genen­tech group just re­cent­ly pledg­ing up to $1.8 bil­lion to the team’s hunt for a bet­ter cap­sid. As part of the deal signed in Oc­to­ber, Dyno went to work with the Spark team at Roche build­ing bet­ter mod­els of pro­to­type AAV vec­tors and look­ing to over­come some of the bar­ri­ers that have kept the ther­a­py’s po­ten­tial cor­ralled to a lim­it­ed set of or­gans.

The Roche deal came on the heels of sim­i­lar deals with Sarep­ta and No­var­tis, both of which looked to har­ness the com­pu­ta­tion­al pow­er be­hind Dyno’s de­signs.

Tar­get­ing a Po­ten­tial Vul­ner­a­bil­i­ty of Cer­tain Can­cers with DNA Dam­age Re­sponse

Every individual’s DNA is unique, and because of this, every patient responds differently to disease and treatment. It is astonishing how four tiny building blocks of our DNA – A, T, C, G – dictate our health, disease, and how we age.

The tricky thing about DNA is that it is constantly exposed to damage by sources such as ultraviolet light, certain chemicals, toxins, and even natural biochemical processes inside our cells.¹ If ignored, DNA damage will accumulate in replicating cells, giving rise to mutations that can lead to premature aging, cancer, and other diseases.

Tom Barnes (Orna)

The mR­NA era is here. MPM be­lieves the fu­ture be­longs to oR­NA — and Big Phar­ma wants a seat at the ta­ble

If the ultra-fast clinical development of Covid-19 vaccines opened the world’s eyes to the promises of messenger RNA, the subsequent delays in supply offered a crash course on the ultra-complex process of producing them. Even before the formulation and fill-finish steps, mRNA is the precious end product from an arduous journey involving enzyme-aided transcription, modification and purification.

For Bristol Myers Squibb, Novartis Institutes for Biomedical Research, Gilead’s Kite and Astellas, it’s time to rethink the way therapeutic RNA is engineered.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 102,000+ biopharma pros reading Endpoints daily — and it's free.

Fol­low biotechs go­ing pub­lic with the End­points News IPO Track­er

The Endpoints News team is continuing to track IPO filings for 2021, and we’ve designed a new tracker page for the effort.

Check it out here: Biopharma IPOs 2021 from Endpoints News

You’ll be able to find all the biotechs that have filed and priced so far this year, sortable by quarter and listed by newest first. As of the time of publishing on Feb. 25, there have already been 16 biotechs debuting on Nasdaq so far this year, with an additional four having filed their S-1 paperwork.

Steve Cutler, Icon CEO (Icon)

In the biggest CRO takeover in years, Icon doles out $12B for PRA Health Sci­ences to fo­cus on de­cen­tral­ized clin­i­cal work

Contract research M&A had a healthy run in recent years before recently petering out. But with the market ripe for a big buyout and the Covid-19 pandemic emphasizing the importance of decentralized trials, Wednesday saw a tectonic shift in the CRO world.

Icon, the Dublin-based CRO, will acquire PRA Health Sciences for $12 billion in a move that will shake up the highest rungs of a fragmented market. The merger would combine the 5th- and 6th-largest CROs by 2020 revenue, according to Icon, and the merger will set the newco up to be the second-largest global CRO behind only IQVIA.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 102,000+ biopharma pros reading Endpoints daily — and it's free.

S&P ex­pects steady ero­sion in Big Phar­ma's cred­it pro­file in 2021 as new M&A deals roll in — but don't un­der­es­ti­mate their un­der­ly­ing strength

S&P Global has taken a look at the dominant forces shaping the pharma market and come to the conclusion that there will be more downgrades than upgrades in 2021 — the 8th straight year of steady decline.

But it’s not all bad news. Some things are looking up, and there’s still plenty of money to be made in an industry that enjoys a 30% to 40% profit margin, once you factor in steep R&D expenses.

Tal Zaks, Moderna CMO (AP Photo/Rodrique Ngowi, via still image from video)

CMO Tal Zaks bids Mod­er­na a sur­prise adieu as biotech projects $18.4B in rev­enue, plots post-Covid ex­pan­sion

How do you exit a company after six years in style? Developing one of the most lucrative and life-saving products in pharma history is probably not the worst way to go.

Tal Zaks, Moderna’s CMO since 2015, will leave the mRNA biotech in September, the biotech disclosed in their annual report this morning. The company has already retained the recruitment firm Russell Reynolds to find a replacement.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 102,000+ biopharma pros reading Endpoints daily — and it's free.

Ken Frazier, Merck CEO (Bess Adler/Bloomberg via Getty Images)

UP­DAT­ED: Mer­ck takes a swing at the IL-2 puz­zle­box with a $1.85B play for buzzy Pan­dion and its au­toim­mune hope­fuls

When Roger Perlmutter bid farewell to Merck late last year, the drugmaker perhaps best known now for sales giant Keytruda signaled its intent to take a swing at early-stage novelty with the appointment of discovery head Dean Li. Now, Merck is signing a decent-sized check to bring an IL-2 moonshot into the fold.

Merck will shell out roughly $1.85 billion for Pandion Pharmaceuticals, a biotech hoping to gin up regulatory T cells (Tregs) to treat a range of autoimmune disorders, the drugmaker said Thursday.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 102,000+ biopharma pros reading Endpoints daily — and it's free.

Glax­o­SmithK­line re­thinks strat­e­gy for Covid-19 an­ti­body — not the Vir ones — af­ter tri­al flop. Is there hope in high-risk pa­tients?

In the search for a better Covid-19 therapeutic, GlaxoSmithKline and Vir have partnered up on two antibodies they hope have a chance. GSK is also testing its own in-house antibody, and early results may have shut the door on its widespread use.

A combination of GSK’s monoclonal antibody otilimab plus standard of care couldn’t best standard of care alone in preventing death and respiratory failure in hospitalized Covid-19 patients after 28 days, according to data from the Phase IIa OSCAR study unveiled Thursday.

Photo: Shutterstock

Bio­phar­ma's suc­cess rate in bring­ing drugs to mar­ket has long been abysmal. Can new tools help rewrite that trou­bled past?

In 2011, a team of researchers at British drugmaker AstraZeneca had a problem they were looking to solve.

For years, drug discovery and development were a wasteland for innovation. Novel drugs largely fell into one of two categories — monoclonal antibodies and small molecules — and new therapeutic modalities were hard to come by. After a rush of promising approvals in the late 1990s — including then-Biogen’s CD20 targeting antibody breakthrough Rituxan — the field stagnated and attrition rates stayed sky-high. What exactly is the industry doing wrong? AstraZeneca asked itself.

Endpoints Premium

Premium subscription required

Unlock this article along with other benefits by subscribing to one of our paid plans.