Showing posts with label Big Data. Show all posts
Showing posts with label Big Data. Show all posts

Monday, December 22, 2014

Big Idea 2015: Haute Apothicaire




There are medicines & then there is Haute Apothicaire
From what’s written on the wall, 2015 could well be the year when good old medicine finally goes designer chic. Even if not for the masses just yet, a privileged few could soon have their meds custom made to match their genes - well, just as soon as the fitting trials are done & the master says aye!
For those who didn't quite get my pun, I am indeed talking about personalized medicine going mainstream in 2015 - From Novartis making the first move to bring protocol to patient through its series of #SIGNATURE trials & JnJ opening up its clinical data vaults to Yale for open research to #Google #Calico aiming to tackleageing through big data & bigger collaborations, it's clear there's this huge trend of big pharma going all guns blazing after the alluring promise of personalized medicine.
This alliance with big data is in some way an existential manoeuvre by big pharma poised precariously on the patent-cliff & thus most initial approvals of designer drugs would be those that have been pulled up from the gorges of failed/failing PII/ PIII trials. Either which way, there is no doubt that the best use of cutting-edge drug design capability is only when it is coupled with the invaluable human genomic & proteomic data decoded in the past decade and using the #bigdata analytics to throw up the statistically best therapeutic match for an individual patient.
If one goes by the regularity with which FDA has been approving companion diagnostics and genetic tests over past few years; the apparent merit they see in the label of an approved drug referencing the genomic biomarker(s) & the pragmatic use of conditional approvals such as the recent one for Translarna by the EMA, it appears the regulators too are in sync with this new reality & have been gearing upto deal with the potential multitudes of personalized therapeutic applications & evolve an approval mechanism that's applicable despite the indication not being orphan enough.
Even as connoisseurs may not quite like it, high fashion attains scale only when it goes mainstream & by analogy the designer drugs too would find their way to the masses eventually, I only hope sooner & safer.
#BigIdeas2015 #Google #Calico #Bigdata

Thursday, September 11, 2014

Forget ageing, let Google manage you!


I can visualize the oodles of intellectual optimism and excitement generated when Bill Maris & Ray Kurzweil sat down for the very first time to envision Calico to ‘Tackle Ageing’ & prolong lives. But despite an equally bold mission statement saying Calico will ‘Devise interventions that enable people to lead longer and healthier lives’, I find it difficult to believe that Larry would’ve agreed to sink millions in this seemingly altruistic goal of creating a clutch of wonder pills that increase human lifespan, in about a decade plus, without fully addressing some key questions, most importantly why Google, why ageing.
Even if in their new-covert enthusiasm (to healthcare), Bill, Larry et al would’ve down-played the primary question of why Google, it’s unlikely a battle-worn pharma baron like Richard Gonzalez would ignore the same. But the fact that Gonzalez is firmly sold on to this idea now indicates that Calico’s actual potential is not definitely limited to mere (& slow) fruition of their stated long-term vision.
Secondly, while it may look like a really compelling cause given the promise of long-life, scientifically I’m still unsure what makes ageing research an attractive business proposition considering Alzheimer’s drug development is struggling with a depressingly high failure rate of 99.6%? Perhaps Google saw some merit is in this very realization that ‘Alzheimer’s Cure’ is pretty much an oxymoron and prevention is where the hope lies. Of course ageing is much more than Alzheimer’s and hopefully the rest of the pathways that arrest degeneration don’t have as many pitfalls and mines as Alzheimer’s does, but how is this untested promise of tackling ageing alluring enough for a healthcare upstart like Google to jump-in with a big-ticket investment?
Through my freestyle inquisitive ramblings below, I have tried to better understand this stealth start-up & see if my above hypothesis makes some sense. Also, despite knowing CALICO is set-up as an independent company, I have dealt with this like Google’s still the boss.

What’s in a name? – Quite a bit actually

Other than the obvious fact that it’s short for ‘California Life Company’, the name surely reminds one of a Calico cat. Art Levinson was apparently okay with the cat connotation in connection with a cat’s nine lives (read: longevity). But looking at all the illustrious names on the team, one'd surely suspect the connotation is far deeper than it looks & there's more to Calico than the cat itself.
After some serious searching, I could make some connection. The biological background to the curious phenomenon called calico cats lies in a spontaneous genetic mechanism called X-inactivation & the effects associated with relative Degree of Skewing (DS) of paternal & maternal X chromosomes among female of the species.The most interesting find in this connection is a study published on NCBI that establishes a link between higher DS & longer lifespan – one of the authors of this paper incidentally is Prof. Kaare Christensen of Danish Ageing Research Centre, University of Southern Denmark.
Great conjectures all, but I’m wondering where does this leave male-longevity? :-)

Let me get this off my list

Is Calico Google Health 2.0? Google Health (GH) was merely a medical data management initiative, a natural offshoot of Google’s cloud aspirations but also Google’s first tentative entry into healthcare. As a “personal health information centralization” service, GH depended on people adapting the same and needed other support organizations to achieve scale. While Google never claimed they did, GH’s monetization was perhaps through positioning healthcare products, solutions to individual users based on the information available against their profiles. Given this background, it doesn't look likely Calico is just rehashed Google Health.
Now the primary reason for closing GH is stated to be ‘the low adoption’ of the platform by users - Really? Did Google really expect to cover the world within one year after it released the finalized version of GH in 2010 before announcing its retirement in 2011? This super-short life-cycle makes me suspect if Google Health was a recce for Calico if not its previous version. If indeed GH was planned as a recce or even if it was coincidental, I guess there would have been some key takeaways; learning from this escapade that went into design of Calico – a few that come to mind are;
  • What healthcare data an average individual is comfortable putting up on cloud & what she/ he isn’t
  • How easy or tough it is to ensure consistent updating of good quality, verifiable data?
  • The minimal data points required to estimate an undisclosed indication(s) (user needs) and have a firm basis for monetizing the venture.-------- E.g.when someone uploads Vitamin-D levels on a quarterly basis, other than the fact that the subject is Vit-D deficient if), estimating the manifesting condition is not still possible unless some other kind of data is uploaded which when read together with the first throws up a tangible story. Now if the same user uploads data on bone-density and say a Vertebral CT Scan, it may be safe to assume the user’s actual condition is that of spondylosis pain & thus attempt reaching out to the user and propose solutions.
  • Did the solutions advised/ advertised, push-notification made to the users translate to any tangible change in their choices (of where/ what to spend) & what percentage of users were inclined to take endorsements as ‘real’ health solutions?
It’d be reasonable to assume that Google Health conceived in 2008 was piggy backing on the quantified-self trends that incidentally are still strong and was essentially designed to accept, process data fed manually either by user or by supporting stakeholder organizations. However given the proliferation of the wearable digital health feedback systems in past 2 years & given the inherent limitations of the GH platform to integrate internet of things may have prompted Google to discontinue GH and think of a completely new to way pursue its Health ambitions.

Why is Google thinking health?

Calico in its stated vision is unabashedly a healthcare company & nowhere hints at anything even remotely Googlish (sic). So what prompted Google to enter a domain that’s clearly & completely unconnected to its core-competencies. If I let alone the self-actualization needs of Larry & Sergei and Google’s avowed itch for innovation & paradigm-setting, what unstated and unassailable advantages Google already has/ will have over other Pharma biggies who've been there and done but yet can’t confidently talk of prolonging lifespan?
Of course Sky Nash on LBB earlier did say ‘Utilizing the world’s leading minds and Google’s raw computational power, Calico promises to lift, correct and protect humans from aging’. Computational power is fine, but Google neither owns nor has even an indirect stake in any of the invaluable clinical data generated by pharmaceutical innovator organizations over thousands of clinical failures and some successes. So what other kind of data can potentially help them do a better job of putting out drugs and those that tackle ageing at that? – could it be;
  • Not to the same level of detail as FB has and not as structured, but Google has user information that could be put to use in wellness business.
  • Google’s probable access to the data of online sequencing companies like 23andMe & DNANexus (GV invested into these companies).
If user data is at the base of this opportunity, how is Google data better than FB's which is far better differentiated & spread across billions of users? - I guess the crux lies in the data semantics. The semantics of a structured data query peppered with specifics generated from mapping user preferences through “Likes” (as for FB) V/s of an unstructured, spontaneous search history based data (as for Google) can be greatly different.
However given that Google search is more personal & private than activity on FB account (contestable, as the lines are blurring ever since smart phones with ‘location-on’ took over) people are likely less inhibited in what they search for & this inherently makes their profiling based on search history a lot more accurate and actionable at a one-one level. Whereas on FB, the likes by the individual’s at best indicate a broad interest rather than an individual trait and importantly don’t necessarily indicate what the user is looking for – Let me just say that the data Google has access to & generates every second is more valuable than what FB would have when it comes to identifying a heath consumer's need!
Example: consider & compare the following Google search keyword composition, as recorded from a single advanced search or from a search history over a period of time…
  • 40ys+heart disease+arrhythmia+female+supplements?+self-diagnosis+vitamins+non-surgical treatment+jogging+yoga+meditation+fenofibrate+non-surgical+hereditary?+childhood eosinophilia+zinc+PCOS+India+traction safety+physiotherapy
With the following FB profile+activity
  • 40yrs+female+1500 friends across 20 countries+Likes “Yoga”; “Herbalife”; “working women” “Titanic the Movie”; “Savage Garden” “feel” worried, bored, depressed variably across a month & with status-updates that dealt with topics & images that demonstrate ‘feeling generally great”; “euphoria”; “health problems”; “heartbreak’; “depression” et al
When it comes to how much one can profile the user based on the above input, I feel Google search history wins hand down on the degree of specificity of data & for the sheer number of conjectures and probabilities possible. FB data on the other hand is specific on non-health aspects & even if an effort is made to interpret the semantics, it’s likely to end up with ‘probably-true-observations’ but not much actionable observations on aspects related to health. So as long as health info is considered ‘stealth info’ by users, Google is where people are going to end-up.
Also, with the big pharma companies increasingly open about sharing their clinical data (Ref: JnJ tying up with Yale university) with an intention to maximizing the data’s monetization potential, Google may find willing partners in them and thus access to clinical data may not remain a weakness anymore for Calico.

Does Google Ventures hold a clue to Google’s health ambitions?

Yes, it appears. Google Ventures (GV) portfolio does look representative of the arguments put forth earlier on why Google is into Health? The life science portfolio is populated with companies that have a definite personalized medicine/ health focus --à 23andMe (personal genetics); one medical group (personal medicine); foundation medicine (cancer diagnostics, genomics analysis); DNANexus (cloud based DNA sequencing); Flatiron health (Onco data platform); Predilytics (advanced healthcare analytics); Transcriptic (next generation lab automation); SynapDX (blood diagnostics); Rani Therapeutics (oral large molecule biotherapeutics); Doctors on Demand; AdiMAb (Yeast based antibody discovery platform). Other than life Sciences, GVs investment into upstart companies leveraging Hadoop for healthcare data management also is relevant to the personalized medicine, E.g. Cloudera; Clearstorydata etc.
As the trends indicate, it is imminent Google Ventures will start investing into companies working on wearable/ injectable health devices, such as the one by Proteus and obtain a stake in the high value vital body data domain? But once cannot ignore the fact that GV is absent from most of the big wearable start-up funding deals in 2014 namely, NanoHealth ($135 million), Alignment Healthcare ($125 million), Proteus ($120 million), MedHOK ($78 million), Lumeris ($71 million), Zenefits ($67 million), and Doximity ($54 million)) – This reluctance is surprising considering Google otherwise appears bullish on health.
Update 15 Sep 2014Though it isn't linked to Google Ventures, I did miss out including Google's smart contact project for which they have tied-up with Novartis - While at this stage it isn't clear the lens can send up the date to the mobile phone, assuming its on the cards, this could be a cutting edge usage of internet of things in medical devices. Similarly the recent Google acquisition of Lift technologies: while Lift spoon is indeed a amazingly simplistic and utilitarian device that improves basic quality of living for those effected by Parkinson's, I guess the Lift Pulse & Lift Stride technologies are what are going to help Google integrate this company into its digi-health future.

Data + User engagement + Alliances + Tailored drugs = Longevity

From the above hypothesis & conjectures, it's safe to infer that Calico would pursue its vision of tackling ageing through its mission to engineer a large-scale, data driven personalized medicine revolution.
The various nuts and bolts that’ll naturally be part of this greater mission could include aspects like; Enabling users to get into the quantified-self mode & leverage this conversion to create a data-set that could be utilized for genetic profiling & drug discovery. Using the personal charisma of its start-studded team to cobble together alliances with other major pharmaceutical innovator organizations like Abbvie, get access to their idle clinical data & maximize the utility of the same by enabling open sourced research & thus create an ecosystem of mutually dependent pharmaceutical innovation clusters.
The leadership at Calico looks well suited for the purpose and their individual strengths in sync with the above vision & mission – Below is a quick peep into what individual value-add they’ll potentially get to the table.
  • Art Levinson: Innovator, administrator par excellence with a jaw dropping professional pedigree – Unquestionably most useful in getting alliance partners from pharma & technology domains.
  • Hal V Barron: Knows how stuff works or not in clinic, what pitfalls exist in clinical trial & just how much of it can be predicted genetically – Deep domain knowledge of another lifespan linked disease domain of Cardio Vascular therapeutics.
  • David Botstein: The one who’ll know what to do with big data on genetic ancestries from the likes of 23andMe - Can help put out diagnostic tools (mobility driven; home-use; Molecular diagnostics? Companion diagnostics for personalized medicine)
  • Cynthia Kenyon: Ageing expert, figuratively speaking the one who can map out a calico cat - The quintessential tea-taster, the screener of potential drugs that act on individual ageing triggers.
  • Robert Cohen: The one who understands the opposite of longevity, programmed cell death nee apoptosis, the devil’s advocate to Cynthia - Understands cancer triggers, and blocks & individual variances in each genetic make-up

Is Calico’s “challenge to ageing” the new drug-discovery paradigm?

Or is it just a play of words & a smart pitch to the healthcare consumer who hitherto had to make do with rather promise-less solicitations & six-month increases in survival after a year-long therapy?
Whatever may be intention & the outcome, Calico is sure to change the way healthcare is viewed, pursued henceforth and the health consumers are in for an entirely new & enticing solicitation.
Exciting times indeed!
----------------------------------------------------------------------------------------------------------------------
Image courtesy (partly modified) - Wikipedia