Bonnie Berger, MIT professor of mathematics (via MIT)

MIT lab cre­ates tool to find can­cer mu­ta­tion dri­vers 'any­where in the genome'

Pre­ci­sion on­col­o­gy has been her­ald­ed as the way for­ward in treat­ing can­cer, but search­ing for dri­ver mu­ta­tions has com­pli­cat­ed that jour­ney as ex­ist­ing meth­ods can’t be ap­plied to the en­tire genome.

But a new method out of MIT looks to change the way can­cer dri­ver mu­ta­tions are found and, in do­ing so, speed up the process. A group out of Bon­nie Berg­er’s lab at the in­sti­tu­tion has de­vel­oped what they named “Dig,” an in­ter­ac­tive map for find­ing dri­ver el­e­ments and mu­ta­tions “any­where in the genome,” ac­cord­ing to a pa­per pub­lished this week in Na­ture Biotech­nol­o­gy.

The group de­vel­oped a deep learn­ing mod­el by train­ing the AI to map can­cer-spe­cif­ic so­mat­ic mu­ta­tion rates us­ing the Pan-Can­cer Analy­sis of Whole Genomes, or PCAWG, a dataset in­clud­ing 37 can­cer types. The team used high-res­o­lu­tion epi­ge­net­ic as­says from healthy tis­sues to guide the pre­dic­tions.

To dig deep­er, the lab ap­plied Dig to an­oth­er task: find­ing new cod­ing and non-cod­ing can­di­date dri­vers of can­cer us­ing whole-genome, whole-ex­ome and tar­get­ed se­quenc­ing can­cer datasets avail­able to the pub­lic.

Dig ap­peared to beat out many com­pet­ing meth­ods, the group found, not­ing their tool had the high­est mea­sure of ac­cu­ra­cy in 24 out of 32 co­horts in PCAWG. Skin and blood can­cers in PCAWG were ex­clud­ed be­cause of “lo­cal hy­per­mu­ta­tion process­es,” they wrote.

The group’s tool al­so “matched or ex­ceed­ed the per­for­mance” of ex­ist­ing meth­ods used to search ex­cess of mu­ta­tions in dri­ver el­e­ments that had al­ready been iden­ti­fied.

The ac­cu­ra­cy is thanks, in part, to the deep learn­ing tool’s abil­i­ty to find ac­tive tran­scrip­tion start sites and oth­er epi­ge­net­ic struc­tures and then as­so­ciate those struc­tures with the mu­ta­tion rates, the group wrote.

In essence, the com­put­er sci­ence and AI lab led by Berg­er, head of the com­pu­ta­tion and bi­ol­o­gy group, used AI to pre­dict mu­ta­tion rates and com­pared it to re­al da­ta to help train the mod­el to look for mu­ta­tions through­out the genome.

Find­ing dri­ver mu­ta­tions has typ­i­cal­ly been done us­ing ar­bi­trary re­gions of the genome, rather than the broad­er ge­net­ic land­scape, a process that is time-con­sum­ing and ex­pen­sive, the group writes in their pa­per. Pre­vi­ous meth­ods have fo­cused on DNA se­quences that are in­volved in mak­ing pro­teins, like cod­ing and pro­mot­er se­quences, which leave out a big swath of the genome.

“These lim­i­ta­tions con­tribute to cat­a­logs of can­cer dri­ver el­e­ments re­main­ing in­com­plete, par­tic­u­lar­ly in the non-cod­ing genome, hin­der­ing pre­ci­sion on­col­o­gy,” ac­cord­ing to the pa­per.

The group found rare mu­ta­tions re­spon­si­ble for can­cer that oc­cur out­side the al­ready ex­plored re­gions, per­haps shin­ing a light on as much as one-tenth of tu­mors.

“Al­though the dri­ver can­di­dates we re­port — in cryp­tic splice sites, 5′ UTRs and rarely mu­tat­ed genes — oc­curred at low fre­quen­cies in­di­vid­u­al­ly, our es­ti­mates sug­gest that they col­lec­tive­ly con­tribute to the dis­ease pathol­o­gy of up to 10% of tu­mors (sum­ming across the per­cent of tu­mors pre­dict­ed to car­ry ex­cess mu­ta­tions in each of these el­e­ments),” the group wrote in the pa­per.

The team claims the tool can be used quite broad­ly and can churn out re­sults in a mat­ter of min­utes. They have made it pub­licly avail­able.

“Through this frame­work, our maps en­able mil­lions of mu­ta­tions to be eval­u­at­ed in ar­bi­trary can­cer co­horts in min­utes us­ing the re­sources of a per­son­al com­put­er,” the MIT lab wrote in the pa­per.

The Fac­tors Dri­ving a Rapid Evo­lu­tion of Gene & Cell Ther­a­py and CAR-T Clin­i­cal Re­search in APAC

APAC is the fastest growing region globally for cell & gene therapy trials representing more than a third of all cell & gene studies globally, with China leading in the region. 

APAC is the leading location globally for CAR-T trials with China attracting ~60% of all CAR-T trials globally between 2015-2022. The number of CAR-T trials initiated by Western companies has rapidly increased in recent years (current CAGR of about 60%), with multiple targets being explored including CD19, CD20, CD22, BCMA, CD30, CD123, CD33, CD38, and CD138.

The End­points 11; blue­bird's $3M gene ther­a­py; Bio­gen tout new neu­ro da­ta; Harsh re­views for can­cer drugs; and more

Welcome back to Endpoints Weekly, your review of the week’s top biopharma headlines. Want this in your inbox every Saturday morning? Current Endpoints readers can visit their reader profile to add Endpoints Weekly. New to Endpoints? Sign up here.

Reading about John Carroll’s pick of biotech’s most promising startups has become a treasured tradition. If you ever get curious about previous classes of the Endpoints 11, you can find all of them (plus a number of our other regular specials) here.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 150,800+ biopharma pros reading Endpoints daily — and it's free.

EMA warns of short­ages of two Boehringer heart drugs due to a spike in de­mand

The EMA is putting EU member states on alert over the shortage of two drugs that counter heart attacks due to an uptick in demand.

On Friday, the EMA sent out a warning that two Boehringer Ingelheim drugs are experiencing a shortage: Actilyse and Metalyse. The drugs are used as emergency treatments for adults experiencing acute myocardial infarction, or a heart attack, by dissolving blood clots that have formed in the blood vessels.

The End­points 11: The top pri­vate biotechs in pur­suit of new drugs. Push­ing the en­ve­lope with pow­er­ful new tech­nolo­gies

Right around the beginning of the year, we got a close-up look at what happens after a boom ripples through biotech. The crash of life sciences stocks in Q1 was heard around the world.

In the months since, we’ve seen the natural Darwinian down cycle take effect. Reverse mergers made a comeback, with more burned out shells to go public at a time IPOs and road shows are out of favor. And no doubt some of the more recent arrivals on the investing side of the business are finding greener pastures.

Endpoints Premium

Premium subscription required

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

Jim Wells (UCSF)

An­ti­bod­ies once act­ed on­ly as pro­tein block­ers. Now, sci­en­tists are find­ing new ways to make them pro­tein de­stroy­ers

The first lab-made antibody medicine was approved in 1986 — it bound to an antigen known as CD3 on T cells and was meant to prevent kidney transplant rejection. While antibody technology improved, most antibodies were made as blocking agents, neutering clamps that attacked cells and proteins.

But then scientists got creative with their engineering. They made antibody-drug conjugates, or ADCs for short, which attached toxins or drugs to the antibodies, enabling them to kill cells. Then they made CAR-T therapies, which attached a patient’s T cell to the targeting fragment of an antibody, to destroy cancer cells.

Endpoints Premium

Premium subscription required

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

As­traZeneca, Mer­ck cull one Lyn­parza in­di­ca­tion in heav­i­ly pre­treat­ed ovar­i­an can­cer pa­tients

Just one day after blockbuster Lynparza got access to another indication in China, its Big Pharma owners have decided to withdraw it in certain patients after reviewing Phase III data.

The two companies that work together on Lynparza decided to recall one of the indications several weeks ago in a specific type of ovarian cancer, Lynparza’s first indication when it was first FDA-approved in 2014. Initial data showed that rates of overall survival in patients with at least three rounds of chemo before getting on the PARP inhibitor were lower than in patients with less previous chemo treatment.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 150,800+ biopharma pros reading Endpoints daily — and it's free.

Fu­ji­film con­tin­ues CD­MO ex­pan­sion, break­ing ground on $435M UK site

Fujifilm’s CDMO arm, Fujifilm Diosynth, has been on a roll this month as the company has recently broken ground on a major project in Europe and it appears to be keeping up the momentum.

Fujifilm Diosynth announced that it has kicked off an expansion project for its microbial manufacturing facility at its campus in the town of Billingham, UK, in the northeast of England.

The 20,000 square-foot, £400 million ($435 million) expansion will add clean rooms, purification suites and a packing area along with more space for the manufacturing itself.

Solicitor General Elizabeth Prelogar

Should SCO­TUS hear Am­gen's Repatha case? So­lic­i­tor gen­er­al says no

Back in April, Amgen said it was encouraged by the solicitor general’s anticipated review of its Supreme Court petition to rehear a Repatha patent case. They’re likely much less optimistic about the outcome now.

Solicitor General Elizabeth Prelogar wrote in a recent 27-page brief that Amgen’s arguments “lack merit and further review is not warranted.”

The case traces back to a suit filed in 2014 against Sanofi and Regeneron’s Praluent, which ended up beating Amgen’s PCSK9 blockbuster Repatha to market by a month just a year later.

Phil Sharp, Nobel Prize laureate (L), and John Carroll, Endpoints News co-CEO (via Michael Last)

The End­points 11: Fire­side chat with No­bel Prize lau­re­ate Phil Sharp

The following Q&A has been edited for length and clarity.

John Carroll:

We’ve had a chance to talk a little bit before here about some of the things that you’ve done. Just really remarkably, a lot of the things that you’ve done early in your career puts you in the path with some amazing science that has had an absolutely huge impact in terms of what we’re seeing now on drug development and some of the new technologies that are coming out here, and not only the new technologies, but also some of the most remarkable people ever.

Endpoints Premium

Premium subscription required

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