Biowebtronics Biotech, startups, web development and internet of things.

How to use Marlowe in a sharepoint emergency!

Marlowe is a job tracking tool for the mass spectrometry service. It is a little buggy but it is better than submitting jobs by email!!

Quick Start

First you need to create an account with your email and a new password.

Once logged in, go to settings and add the project code where jobs get charged too.

Once that is created you can go ahead to the jobs section.

Click new job.

Chemical formula is case sensitive.

Fill in the details, the molecular weight will be calculated for you.

Hazard buttons don't work so well right now so if you could just put the hazards in the comments that would be awesome!

Click the button to attach a chemical structure. We like jpg's pleasethanks.

Then click Submit.

Note that after clicking submit, it will submit it but you won't know!

Click the jobs button at the top of the page to return to your jobs so see your newly created job!!

That's it, cya!

P.s if problems email Ben Miles, search the IC directory for it.

Attempted valuation by the risk adjusted net present value (rNPV) method

Versartis

This week I wanted to attempt a valuation of a company called Versartis.

Versartis is a single asset company based around a recombinant human growth hormone (rhGH). The key value proposition here is that Versartis can extend the half-life of hGH in circulation somewhere around 60x, they are aiming for a once-a-month injection for treating pediatric growth hormone deficiency. This is achieved by creating a fusion peptide of hGH and a peptide licensed under the name XTEN from Amunix (from where Versartis was spun-out). Hence the product is an XTENylated hGH known as VRS-317 primarily positioned at treating paediatric human growth hormone deficiency. Improved clinical outcomes are intended to come from improved patient compliance as ultimately it will mean less uncomfortable sub-cutaneous injections for the kids, and extend the lifetime of hGH within a therapeutic concentration window.

Existing hGH drugs on the market are hardly differentiated hence there is an opportunity here to separate from the rest of the market where all the players are present. These include Nutropin (Genentech), Humatrope (Lilly), Genotropin (Pfizer) and Norditropin (Novo Nordisk).

Versartis IPO'd on March 21 2014 on NASDAQ and have previously raised capital all the way through to a series E round!

Risk adjusted Net-Present-Value

I was looking around for a method of valuing a drug co. pre-market and found an article in Nature Biotechnology on a technique which takes into account the risk and time to getting a drug to market and ultimately the revenues when at market and calculates an adjusted value for the present time. The supplementary information for the article comes with a handy spreadsheet model so I hacked away at that for a few evenings.

So here are the parameters I used for calculating the rNPV for Versartis:

Market Data

MARKET DATA
Number of Cases Forecast for Year 1 100,000
Annual Population Growth 0.27%
Peak Market Penetration 45.0%
Revenue Per Unit $30,000
Market Ramp Time to Peak Penetration (Years) 5
Orphan Drug (< 200,000 U.S.)? (y/n) Y
Bottom up annual paediatric GHD cases

Cases Year 1 based on 1:3480 children in North America are estimated to suffer from GHD1.

42.66M children in North America2 + 30.8M children in Europe3 + Japan 16.38M. (population data from Wolfram Alpha.)

Annual number of cases of Paediatric GHD: 25,816.

Top down annual GHD cases

Another way to calculate number of cases per year is to take the annual dollar size of the market ~$3b and divide that by the annual cost of treatment, approximately ~$30,000. Whilst not really from first principles, this takes in account all other applications of hGH which might not nescessarily be a good idea...

Annual number of cases of GHD: 100,000.

Peak market penetration is estimated to be 45% due to how this drug is so differentiated from all others on the market and the assitance it should receive from being an orhan drug.

Revenue per unit is based on the estimated annual cost of treatment per child in the U.S being in the range of $10,000 to $40,000.

Market ramp to peak penetration is estimated to be 5 which I believe is quite short however with orphan drug status the EMA will aid in facilitating the promotion of the drug.

VRS-317 was awarded orphan drug status by the European Medicines Agency (EMA) on September 13, 2013 and the FDA on December 13, 2013.

Trial Information

Patients

Actual enrolment should only be taken from the published data from the clinical trial. Apart from this estimated enrolment is taken from the Clinical Trials database.

Phase Patients Enrolled
Phase 1 504
Phase 2 48 (estimated)5
Phase 3 250 (estimated)6

Time Scales

Phase duration / years
Preclinical Passed
Phase 1 Passed
Phase 2 15
Phase 3 36
FDA 1
Total 5

FDA clearance duration is estimated at 1 year for a fast-track therapeutic.

Versartis completed a Phase 1 trial in 50 adults earlier this year.

A phase 1b/2a pediatric trial is currently underway assessing 6 month growth velocities of kids with pediatric GHD5 due for completion in June 2014, therefore the the lowest duration of 1 year is used..

Additionally there is a planned phase 3 open label extension study allotted for 4 years to assess the long term effects of VRS-3176. However Results may be announced earlier hence an estimate of 3 years.

Costs

Item Cost / $ Notes
Annual Pre-Market Patent Fees $10,000
Annual Preclinical Costs $-
Per Patient Phase 1 $12,000 Hollister-Stier
Per Patient Phase 2 $12,000 Hollister-Stier
Per Patient Phase 3 $6,000 Hollister-Stier
Approval Costs $1,300,000 $309,647 for PDUFA and $500,000-1,500,000+ for NDA preparation (ProPharma Partners; Covance)
Animal studies supporting Phase 1 $500,000 SEC documents referencing 2001 survey
Animal studies supporting Phase 2 $1,000,000 SEC documents referencing 2001 survey
Animal studies supporting Phase 3 $1,500,000 SEC documents referencing 2001 survey
Manufacturing/Marketing Costs + Markup 25%

Not sure about any of these costs, stuck with defaults. Went for 25% marketing costs as they can do the production in bacteria therefore getting a great margin.

Rates

Item Rate Notes
Royalty Rate 5.0%
Discount Rate 20% VC and pharma IRR (cost of capital for biotech)

Risk Mitigated

Phase Risk mitigation entering Phase Notes
Preclinical 0% Completed
Phase 1 0% Completed
Phase 2 66% Ongoing
Phase 3 67% Pharmaceutical Manufacturing and Research Association
FDA 81% FDA FY 2000 Performance Report to Congress for the Prescription Drug User Fee Act

Risk mitigation estimates

I seriously have no idea how to place a percent value on risk mitigation. I am working with 66% and 67% mitigation for phase 2 and phase 3 respectively.

I gave phase 2 a fairly high level of risk mitigation because essentially the in vitro work with XTEN has demonstrated the ability to extend the therapeutic half-life7, needless to say this is probably horribly inaccurate.

Phase 3 has the default from the spread sheet which is an average from the 'Pharmaceutical and Manufacturing Association'.

Results of the model

Metric $
PV of Revenue 1,191,678,737
NPV Cash Flow 831,127,445
Risk-Adjusted NPV 547,557,840

The rNPV comes out at $547.5M however if we look at the current market cap of Versartis we see a value $659.21M (13 May 2014). Therefore it is 17% overpriced by the stock market and should be $22.65 per share.

Patient population vs rNPC

Patient Prevalence Cases Y1 rNPV / $ Notes
1:3480 25,816 139,412,886 Paediatric GHD only, based on epidemiological study and census data.
NA 100,000 547,557,840 All applications of GHD, based on market size and cost of treatment.

You can see from the above table that the only way one can justify the current stock price of Versartis is that the market size is based on prior global, annual hGH sales rather than purely the incidence of paediatric GHD. This is confusing because most of Versartis' PR and clinical trials are aiming at paediatric GHD. Perhaps the market cap is based on if Versartis advanced VRS-317 to adult treatments as well?

Some take away questions

  • How do you estimate risk mitigation of a clinical trial phase?
  • Should the rNPV be equal to the market cap if the stock is accurately priced?
  • How do you estimate market penetration of a product?
  • Why does this model not take into account cash on the books and burn rate? Surely the length of runway the company has must play some role in defining value?
  • Is Versartis being valued based on conditions beyond paediatric GHD?

Conclusions

I have learnt it's not just about having a great valuation model. You also need models to determine the risk mitigation and market penetration.

I would be able to provide more accurate costs if I had access to a full quarterly report, but Versartis haven't been public long enough to file that information yet. Additionally I think I will need to play with this model to accomodate different cost types. Additionally I did not have good data for manufacturing and marketing costs.

Whilst I said the market is overvaluing the stock by 17% that's a difference that would easily fall in to an error caused by an in correct risk mitigation rate or time scale for phase III completion. So I can't say with any certainty how good the market price is.

I have to say that I was optimising parameters to get the rNPV to fall into a ball park of the market cap which is probably a stupid idea...

I in no way attest to this being a good analysis. However it is my first attempt so I wanted to learn and keep learning. So get in touch with me on twitter at @BenNMiles and call me out on my BS numbers.

I'm looking forward to seeing the results of the Phase 2a trial in mid June 2014.

Disclaimer

I have no idea what I'm doing so you shouldn't make any investment decisions based on anything I write on my blog. Don't hold me responsible for any financial losses or even gains. I'm doing this for fun not to inform investment decisions.

References

  1. http://www.jpeds.com/article/S0022-3476(94)70117-2/abstract.
  2. http://www.wolframalpha.com/input/?i=population+of+u.s+aged+3-11+years
  3. http://www.wolframalpha.com/input/?i=population+of+europe+aged+3-11+years
  4. Yuen, K., Conway, G. S., & Popovic, V. (2013). A Long-Acting Human Growth Hormone With Delayed Clearance (VRS-317): Results of a Double-Blind, Placebo-Controlled, Single Ascending Dose Study in Growth …. The Journal of …, 98(6), 2595–2603. doi:10.1210/jc.2013-1437
  5. http://clinicaltrials.gov/ct2/show/NCT01718041?term=versartis&rank=3
  6. http://clinicaltrials.gov/ct2/show/NCT02068521?term=versartis&rank=1
  7. http://www.amunix.com/content/technology/index.htm

Maybe you should hire yourself?

As a final year PhD student there is a question Ihear a lot “So what are you doing after?”. The majority of answers I hear are either “I don’t know” or “Hmmm maybe a post-doc?”. And I think these are perfectly acceptable answers. However I think a more important question is “What are you passionate about?” as this shifts the focus from just getting a job, to finding a calling that will satisfy you.

In my opinion you should seek a position that will allow you to achieve your long term goals. Think about where you could work that would best implement your and experience and skills to maximum effect on obtaining that shared goal. Where will your skill set be most leveraged? If there isn’t a company already out here that you could align yourself with to achieve your goal it might just be time to do it yourself.

Over the past year I have been running a small microfluidics business and helping to develop the entrepreneurship scene at Imperial College London through the Imperial Create Lab. So I have a bit of experience of what goes into trying to start and run a business whilst you are studying and I'd like to share my thoughts.

The timing will never be perfect just do it now

Starting a company sounds scary, but then again so does the prospect of doing a PhD. But you never deliver on the final objective in one go. Much like a PhD project, the building and running of a company is broken down into manageable chunks, where your efforts are focussed to reach certain mile stones and then you move on to the next stage. Sure it’s tough but everything, worth doing is.

Starting a company is a little different now. Things are done a ‘lean’ way where risk to capital and time wasting is minimised through formulation of particular assumptions about a business model and assessment of these assumptions by performing rigorous experiments, sound familiar? It should, this is the scientific method applied to running a startup.

As a student you are time poor, however you are surrounded by abundant talent and resources. So you have

The first objective is to identify an appealing product-market fit. There are so many great resources for this so I won’t go over it in too much detail, however the objective is to have your technology developed enough to that it provides value to the customer. Once this value proposition is a established potential customers need to be consulted on your value proposition. Assimilate what you learn from these people and refine your technology and value proposition.

Ultimately if you have solved something no one is willing to pay for it is time to move on to your next idea. By speaking to the customer early and often you can establish that someone won’t pay you for your product before you go ahead and order a few shipping container-fulls! When you bounce your idea off of the market it allows you to better create a solution to a problem from your technology that you have developed.

Why scientists have the makings of great entrepreneurs

There are two reasons I think scientists make great entrepreneurs. Firstly we are well experienced in dealing with failure. It is very rare that your first business will be your most successful one just like how you’ll never nail your first experiment. We understand that there is a cycle of refinement that processes must go through to be the best they can.

Secondly I feel scientists understand the temporary role that assumptions fulfil. Assumptions are critical to make progress however at some point these need to be tested. Business models rely on assumptions like who the potential customer segments will be, at some point this needs to be tested. People need to be spoken to and data needs to be collected to change that assumption into fact. Once all your business model assumptions have been tested you will have a robust business model that puts you in a favourable position to build a successful company and will also look great if you ever wish to pursue investment.

Impact. People. Challenge. Why start ups are amazing

Some of the biggest change agents in our time are companies, not governments and it seems that the best way to change the world is to do it through business. You can’t just start a government to bring about change however you can just start a business.

For me the most exciting thing about start ups are the people. There is something about building a business that attracts visionary people with huge amounts of drive and these qualities are infectious.

From what I have observed helping students start companies is that you get a coming together of some extremely passionate and intelligent people who want to solve a problem or bring delight to people around the world. They see something that annoys them and try and fix it and a startup allows them to solve that problem for everyone else as well.

There is no comfort zone. In a startup you will be constantly pushed to do things you haven’t done before and you will make mistakes but you will learn and grow. It is such a fantastic challenge. Balancing the early stages of a start up is a challenge whilst doing a PhD but the type of skills you will be exercising are a fantastic complement to your usual day to day research.

There is nothing more rewarding than assembling a team to work solely on brining a concept to the market to change the world.

Why university is a great place to build a start up.

The awesome thing about creating a start up at university is the quantity of accessible resources around you. James Taylor who occasionally writes for the Nature Biotechnology blog coined the term ‘The Leverage Startup’. Which is essentially a startup that massively leverages academic resources to minimise risk. I myself have used and I also encourage other students to make the most of:

  • R&D infrastructure
  • Technical Expertise
  • Commercialization resources such as tech transfer offices
  • Non-dilutive funding sources such as government grants
  • Student Entrepreneurship competitions
  • and in some cases pretty favourable intellectual property rights for the student.

From an external point of view these resources usually come with an extremely high price tag, however internally you can speak to a knowledgable academic for the price of a coffee and friendly email. Non-dilutive funding is a great way to advance your product without having to give away any of your equity.

Work on side projects with new people

In my experience the pursuit of a PhD feels like an all consuming experience and it can often lead to opportunities being over looked if they don’t link in to the work that you’re doing. I wanted to address this in my PhD so I embarked to try and imbue spirit of Google 20% into my PhD. At Google, employees are encouraged to spend 20% of their time utilising their skill set in a side project. Often at Google these projects are advertised on an internal bulletin board and teams can form around projects.

Early on I worked on a 20% time project related to polymerisation and since then helping out with the Imperial Create Lab became my main 20% time endeavour. Working on these side projects can enrich your time at university allowing you to get a flavour for different kinds of work and interacting with people from different sectors which can potentially end up with you making a connection with someone who can aid in your search for (self) employment.

What now?

Go out and find kindred spirits, other people who want to fix the world. Even if you don’t want to fix the same problem being involved in a community of people with the drive and desire to build and create things is fantastic support. You will be able to share common skills like how to distribute equity or conduct constructive customer interviews. And more importantly you can share your networks to make connections where you need them.

There are already great communities out there however London is a little more spoilt for choice than the rest of the UK. Meetup.com is a great place to start to find people with similar interests in building companies however if you can’t find the community you are not alone, be the one to go out there and start it, be the nucleus that all of these amazing people can form around. You can do it.

  • Reach out to community or build one!
  • Build a network of people who are driven, and want to make a difference.
  • Find people who align with your vision and ask them to join you on your mission to change a part of the world.
  • Leverage the resources in your institution to minimise risk and gain a competitive advantage.
  • These early stages can be done in your 20% time.

Final thoughts and a bit on if it doesn’t work out.

The big take-away here is that it’s not a big deal to start a company. The key is to minimise risk early on by first speaking to potential customers, and leveraging resources around you in the academic environment. Coalesce a passionate team around the problem you want to solve, and work together developing a great product-market fit where people will actually pay for the technology you are working on.

If ultimately the start up never takes off, or you work tirelessly to build a product and ultimate the company has to wrap up, it doesn’t matter. Think back to your PhD days and how every failed experiment helped you improve yourself and your work. Get back on the horse and solve another problem, and wear your badge of experience with pride.

In silico drug discovery Co.s

This is a post of some notes I made in an attempt to aggregate companies where in silico drug discovery is a main part of their business. Most of the text about the companies is straight from the company website.

Many thanks to @simonbayley, @rowan_UK and @rscoffin for the suggestions.

If you have anymore let me know on twitter @bennmiles

Cheers,

-B

Nimbus US

Nimbus Discovery harnesses cutting-edge computational technologies to uncover breakthroughs in small molecule pharmacology. Nimbus focuses on medically important and highly sought-after disease targets that have proven inaccessible to traditional industry approaches. Nimbus’ robust pre-clinical pipeline includes novel agents for the treatment of cancer, metabolic disease and inflammation. Founded in 2009, Nimbus partnered with Schrödinger to invent and apply a physics-based approach that establishes a new standard for rational drug design.

http://www.nimbusdiscovery.com/

Molplex UK

Molplex is a new class of pharmaceutical company, using its innovative technology to identify and optimise novel treatments for infection, cancer and other serious diseases.

Its pipeline of small-molecule therapeutics is driven by our proprietary platform for drug optimisation, Optiplex - a highly scalable system combining automated decision-making with laboratory automation. Optiplex delivers safer and more effective drugs at radically lower cost.

Molplex aims to discover novel treatments quicker and more effectively, improving the health of patients around the globe.

http://www.molplex.com/

Domainex UK

Domainex is a drug discovery company with a reputation for speed and innovation built on an exceptional track record of drug candidate delivery. It has a world class discovery team with an unrivalled track record of 5 Candidate Drugs delivered in 5 years.

Domainex reduces industry average drug discovery timelines by as much as 30% through the application of novel proprietary technologies and a highly focussed and integrated approach to medicinal chemistry andcomputational chemistry. These technologies and Domainex’s approach also enable it to successfully tackle a greater range of drug target classes, such as kinases, proteases, ion channels, proteins involved in epigenetics and protein-protein interactions.

Domainex has applied its unique technologies and focussed discovery approach to enable it to develop its own pipeline of oncology drugs, including inhibitors of the kinases IKKε/TBK1 and a number of epigenetics related lysine methyltransferases.

http://www.domainex.co.uk/

Biofocus UK

Collaborating with BioFocus' computational team puts leading informatics, molecular modeling and QSAR expertise at your fingertips. Working closely with our colleagues in screening, structural biology, FBDD, medicinal chemistry and with a leading position in the design of our fragment and SoftFocus libraries, our capabilities are available as part of hit identification, hit expansion, hit-to-lead and lead optimization programs.

In addition to contributing to client programs carried out at BioFocus we are able to provide computational chemistry consultancy to troubleshoot and contribute fresh ideas to your projects.

http://www.biofocus.com/

Evotec Switzerland / 600 employees…

Evotec is a drug discovery alliance and development partnership company focused on rapidly progressing innovative product approaches with leading pharmaceutical and biotechnology companies. We operate worldwide providing the highest quality stand-alone and integrated drug discovery solutions, covering all activities from target-to-clinic. The Company has established a unique position by assembling top-class scientific experts and integrating state-of-the-art technologies as well as substantial experience and expertise in key therapeutic areas including neuroscience, pain, metabolic diseases as well as oncology and inflammation.

http://www.evotec.com/

Cresset Group UK

Cresset’s consultants work with a wide range of therapeutically relevant targets and also in many disease areas in which the biological target is unknown. The majority of these projects involve ligand focused virtual screening.

Virtual screening is an effective means of switching chemical series to identify new intellectual property. It is also a fast and cost effective way of testing several different hypotheses.

http://www.cresset-group.com/

Molecular-weightJS

My web application for the chemistry department is really focused on delivering an enjoyable user experience to the researcher, rather than the usual barely functional, ugly as hell software that we get to use.

With this in mind I wanted the app to calculate the molecular weight of a compound for the user, I think it would be a nice thing to see after you just enter the chemical formula, the app just does a bit of magic and does the sum for you.

Due to the app being javascript all the way through the stack, the code for calculating molecular weight based on a given properly formated chemical formula string looks like so:

mass = {
  "H":  1.00794,
  "He": 4.002602,
  "Li": 6.941,
  "Be": 9.012182,
  "B":  10.811,
  "C":  12.011,
  "N":  14.00674,
  "O":  15.9994,
  "F":  18.9984032,
  "Ne": 20.1797,
  "Na": 22.989768,
  "Mg": 24.3050,
  "Al": 26.981539,
  "Si": 28.0855,
  "P":  30.973762,
  "S":  32.066,
  "Cl": 35.4527,
  "Ar": 39.948,
  "K":  39.0983,
  "Ca": 40.078,
  "Sc": 44.955910,
  "Ti": 47.88,
  "V":  50.9415,
  "Cr": 51.9961,
  "Mn": 54.93805,
  "Fe": 55.847,
  "Co": 58.93320,
  "Ni": 58.6934,
  "Cu": 63.546,
  "Zn": 65.39,
  "Ga": 69.723,
  "Ge": 72.61,
  "As": 74.92159,
  "Se": 78.96,
  "Br": 79.904,
  "Kr": 83.80,
  "Rb": 85.4678,
  "Sr": 87.62,
  "Y":  88.90585,
  "Zr": 91.224,
  "Nb": 92.90638,
  "Mo": 95.94,
  "Tc": 98,
  "Ru": 101.07,
  "Rh": 102.90550,
  "Pd": 106.42,
  "Ag": 107.8682,
  "Cd": 112.411,
  "In": 114.82,
  "Sn": 118.710,
  "Sb": 121.757,
  "Te": 127.60,
  "I":  126.90447,
  "Xe": 131.29,
  "Cs": 132.90543,
  "Ba": 137.327,
  "La": 138.9055,
  "Ce": 140.115,
  "Pr": 140.90765,
  "Nd": 144.24,
  "Pm": 145,
  "Sm": 150.36,
  "Eu": 151.965,
  "Gd": 157.25,
  "Tb": 158.92534,
  "Dy": 162.50,
  "Ho": 164.93032,
  "Er": 167.26,
  "Tm": 168.93421,
  "Yb": 173.04,
  "Lu": 174.967,
  "Hf": 178.49,
  "Ta": 180.9479,
  "W":  183.85,
  "Re": 186.207,
  "Os": 190.2,
  "Ir": 192.22,
  "Pt": 195.08,
  "Au": 196.96654,
  "Hg": 200.59,
  "Tl": 204.3833,
  "Pb": 207.2,
  "Bi": 208.98037,
  "Po": 209,
  "At": 210,
  "Rn": 222,
  "Fr": 223,
  "Ra": 226.0254,
  "Ac": 227,
  "Th": 232.0381,
  "Pa": 213.0359,
  "U":  238.0289,
  "Np": 237.0482,
  "Pu": 244,
  "Am": 243,
  "Cm": 247,
  "Bk": 247,
  "Cf": 251,
  "Es": 252,
  "Fm": 257,
  "Md": 258,
  "No": 259,
  "Lr": 260,
  "Rf": 261,
  "Db": 262,
  "Sg": 263,
  "Bh": 262,
  "Hs": 265,
  "Mt": 266,
};

var mW = function (chem) {
    var s = chem.match(/([A-Z][a-z]?)(\d*)/g, chem);
    var compoundWeight = 0;
    for (var i = 0; i < s.length; i++) {
        var element = s[i].match(/([A-Z][a-z]?)/g);
        var count = s[i].match(/([0-9]*)\d/g) || 1;
        compoundWeight += mass[element] * count;
        }
    return compoundWeight
};

It first creates an object of all the atomic masses. The mW function splits up each element and quanity into different strings in an array. Then a loop iterates over each string which extracts the letters to do the look up in the atomic mass object and then it looks for the number to do the multiplication, finally the compundWeight variable is incremented.

You can find the repo on Github if you would like to contribute any changes or improvements.

I like doing code for chemistry and biology. Lets do more!

-B