Days af­ter news leaked that Benev­o­len­tAI had se­cured fund­ing at a rate that would halve its $2 bil­lion val­u­a­tion, the ma­chine learn­ing start­up of­fi­cial­ly an­nounced it raised $90 mil­lion from Sin­ga­pore-based Temasek for a mi­nor­i­ty stake.

The news deals an­oth­er blow to em­bat­tled fund man­ag­er Neil Wood­ford. Benev­o­len­tAI was one of two biotech AI star­tups, along with Ox­ford Nanopore, Wood­ford was re­port­ed­ly prep­ping for auc­tion to raise cash for his still frozen Eq­ui­ty In­come Fund. Nanopore’s $1.5 bil­lion val­u­a­tion re­mains steady, af­firmed by fel­low in­vestor IP Group.

For Wood­ford, the dam­age may have ac­tu­al­ly al­ready been done. Last week, Wood­ford Pa­tient Cap­i­tal Trust an­nounced they were forced to write down the val­ue of one of its as­sets, cost­ing the trust’s net val­ue 4 pence per share and an es­ti­mat­ed $52 mil­lion. The trust has now re­vealed that un­named in­vest­ment is Benev­o­len­tAI.

Nev­er­the­less, Benev­o­len­tAI’s de­flat­ed val­ue may cut in­to the price Wood­ford ul­ti­mate­ly fetch­es for his stake in the com­pa­ny. An FT in­ves­ti­ga­tion from June found Wood­ford’s two firms col­lec­tive­ly owned near­ly a fifth of the Cam­bridge-based biotech. The biotech, in turn, makes up 5,04% of the Eq­ui­ty In­come Fund and, per FT, is Wood­ford Pa­tient Cap­i­tal Trust’s largest sin­gle hold­ing, ac­count­ing for 9.81%.

Wood­ford has said he plans to have enough cash to re­open the Eq­ui­ty In­come Fund by De­cem­ber.

Benev­o­len­tAI’s val­ue rose ex­po­nen­tial­ly af­ter Wood­ford’s ini­tial in­vest­ment in Oc­to­ber, 2014. By its next fund­ing round in 2015, it had reached uni­corn sta­tus, ris­ing from, £190m to £1.16bn, ac­count­ing for 2.86% – near­ly one third – of the com­pa­ny’s 15.9% re­turn that year, ac­cord­ing to num­bers cit­ed by FT. By April 2019, less than two months be­fore Wood­ford was all but forced to sus­pend his fund, it had reached $2 bil­lion.

Benev­o­len­tAI and oth­er ma­chine learn­ing star­tups have been the buzz of biotech the last few years, with their tan­ta­liz­ing promise of cut­ting in­to the steep, now on av­er­age over-$2.5 bil­lion cost it takes to de­vel­op a new drug. Ex­act mod­els vary by com­pa­ny, but broad­ly they promise to cre­ate more tar­get­ed R&D by scan­ning the vast troves of sci­en­tif­ic lit­er­a­ture and mol­e­c­u­lar data­bas­es no hu­man or group of hu­mans could man­age and us­ing it to find pat­terns and de­duce which drugs or com­pounds are most like­ly to be ef­fec­tive. This month, a study from a small­er AI com­pa­ny, In­sil­i­co Med­i­cine, made both in­dus­try-fo­cused and main­stream head­lines when it showed its tech­nol­o­gy could not on­ly scan but al­so in­vent new mol­e­cules tai­lored to tack­le a dis­ease that proved ef­fec­tive in cell cul­tures and mice.

Benev­o­len­tAI ap­pears to have sim­i­lar soft­ware. It dis­tin­guish­es it­self by fo­cus­ing on rare dis­eases and by the re­search fa­cil­i­ty the com­pa­ny ac­quired in Cam­bridge last year, which it says will equip staffers to de­vel­op the drugs from iden­ti­fi­ca­tion by AI sys­tems to clin­i­cal re­search.

The com­pa­ny has part­ner­ships with As­traZeneca and No­var­tis to de­vel­op drugs for chron­ic kid­ney dis­ease and on­col­o­gy. In a state­ment, the com­pa­ny said the cash in­fu­sion will al­low it to “scale” its plat­form for dis­cov­ery and de­vel­op­ment and those col­lab­o­ra­tions will bring in “mean­ing­ful rev­enue” over the next few years.

So­cial im­age: Benev­o­len­tAI via YouTube

As a potentially powerful flu season looms, so does a big test for Roche and its new flu drug, Xofluza. The Swiss giant just got a small boost in advance of that test as the FDA expanded Xofluza’s indication to include patients at high risk of developing flu-related complications.

Xofluza (baloxavir marboxil) was approved last October in the US, the first landmark flu drug approval in 20 years and a much-needed green light for a company that had watched its leading flu drug Tamiflu get eaten alive by generics. Like its predecessor, the pill offered a reduction in flu symptoms but not a cure.

The first-ever Ebola vaccine is on the precipice of approval after the European Medicine’s Agency (EMA) backed the Merck product in this week’s roster of recommendations.

The drugmaker $MRK began developing the vaccine, christened Ervebo, during the West African outbreak that occurred between 2014 and 2016, killing more than 11,000.

The current outbreak in the Democratic Republic of Congo (DRC) has shown case fatality rates of approximately 67%, the agency estimated. Earlier this year, the WHO declared the outbreak — which so far has infected more than 3,000 people — a public health emergency of international concern.

Medical animation has in recent years become an increasingly important tool for conveying niche information to a varied audience, particularly to those audiences without expertise in the specialist area. Science programmes today, for example, have moved from the piece-to-camera of the university professor explaining how a complex disease mechanism works, to actually showing the viewer first-hand what it might look like to shrink ourselves down to the size of an ant’s foot, and travel inside the human body to witness these processes in action. Effectively communicating a complex disease pathophysiology, or the novel mechanism of action of a new drug, can be complex. This is especially difficult when the audience domain knowledge is limited or non-existent. Medical animation can help with this communication challenge in several ways.
Improved accessibility to visualisation
Visualisation is a core component of our ability to understand a concept. Ask 10 people to visualise an apple, and each will come up with a slightly different image, some apples smaller than others, some more round, some with bites taken. Acceptable, you say, we can move on to the next part of the story. Now ask 10 people to visualise how HIV’s capsid protein gets arranged into the hexamers and pentamers that form the viral capsid that holds HIV’s genetic material. This request may pose a challenge even to someone with some virology knowledge, and it is that inability to effectively visualise what is going on that holds us back from fully understanding the rest of the story. So how does medical animation help us to overcome this visualisation challenge?