Why science-focused biotechs should be early adopters of cloud computing
Cloud computing has turbocharged everything from retail to financial services. Now, with managed services, early-stage science-driven biotechs have a golden chance to tap into the advantages of cloud-based computing
Talk of cloud computing is hardly likely to set pulses racing. In terms of excitement, it's probably up there with placing your next stationery order or choosing office furniture. And yet, across everything from financial services to retail, this technology has unleashed dramatic increases in productivity.
Many household names, such as AirBnB or Spotify, were born in the cloud – and would be unthinkable without the elasticity and scalability of computing infrastructure it provides.
Modern-day drug discovery and development, with its high data-processing demands, looks a prime candidate for the benefits of cloud computing. And yet, when it comes to information processing, drug discovery and development remain, for the most part, bewilderingly low-tech. Collaboration on Excel sheets shared in file services remains the modus operandi for many scientists, mirroring the artisanal nature of drug discovery.
However, the advent of managed services in the cloud, has made the benefits of moving into the cloud impossible to ignore. No servers, no hardware - not even of the virtual kind. Instead, you pay for the use of a specific computational resource, whether that’s storing data from high-throughput experiments or analyzing data-heavy 3D structures.
But before we explain why that’s a big deal for early-stage biotechs in particular, let’s clear up a few misconceptions.
Common misconceptions
Why don’t more biotechs invest in cloud-based solutions? It mostly comes down to misunderstandings about complexity, price and intellectual property security.
Misconception 1: Clouds are complex software technology, poorly suited to science-focused biotechs
Drug discovery startups need to be laser-focused on scientific discovery. No early stage biotech wants to be distracted by non-core complexities. Yet this is exactly where managed services trump classic IT infrastructure. They’re easier to implement and easily scalable, with web-based interfaces built for “real people”. Setting up a database, assigning access rights, controlling costs and visualising data can all be done with the click of a button.
This reduces the dependency on in-house IT skills, enabling biotechs to go full steam ahead with the science.
All in all, it’s fair to say that biotechs stand to benefit even more from managed cloud services than tech companies, as they are primarily consumers, rather than builders, of information technology.
Misconception 2: Cloud computing gets expensive once it’s scaled up
The pay-per-use model of cloud services has sometimes been framed as a cost trap, with horror stories of inexperienced users unknowingly racking up massive bills.
Although such examples exist, it’s only really B2B and B2C businesses that find themselves in this situation. They start off small – with a proportional managed services bill – but then grow to the point where millions of customers are using their service. And so the bill grows too.
Biotechs don’t scale according to the number of customers, but with the number of experiments or programs, so costs increase accordingly: step by step. While some biotechs accumulate massive amounts of data (for example imaging), cloud costs can be monitored and managed at the lowest level, down to which assay data is most expensive to handle.
Misconception 3: Data and intellectual property are not safe in the cloud
Data security is, quite rightly, a key concern for biotechs, and the threat level is constantly rising. If you have your own IT infrastructure then the onus is on you to keep those systems secure. You’ll need to be on top of system updates and monitor for emerging threats. That's a challenge enough for most software tech companies, let alone biotech startups!
For managed service providers, cybersecurity is of paramount importance, because any breach could quickly put them out of business. That’s why they have the best people available keeping their systems – and their customers’ data – secure.
When it comes to controlling employees’ access to datasets or apps, all cloud providers are compatible with standard identity management protocols – so if an employee leaves, all access is swiftly revoked.
Having IP-related information scattered across people’s own computers and dropbox creates multiple points of attack – whereas granular access rights and a read/write “digital papertrail” make cloud-based systems more robust. Furthermore, all cloud providers comply with GDPR, HIPAA or similar.
For clinical-stage biotechs, where safeguarding patient data is paramount, this is a vital stamp of approval.
The cloud as a competitive advantage for biotechs
The benefits of cloud computing are not limited to being a lower-maintenance, more modern IT. Managed services can actually be a major enabler for successful R&D, in three key ways.
The benefits of cloud computing are not limited to being a lower-maintenance, more modern IT. Managed services can actually be a major enabler for successful R&D, in three key ways.
Advantage 1: Keep the focus on the science
Cloud-based systems make it simple to set up and scale up your systems – and to access cloud-based datasets. Let’s say you want to make an advanced analysis of the human proteome to filter for interesting targets. OpenTargets data, which is hosted in Google's Cloud Database BigQuery would be a great place to start. And since it is a cloud-managed service, just one click will give your team access to all the data they need
Advantage 2: Safe and convenient collaboration
Drug discovery is perhaps one of the most collaboration-intense endeavors there is. Multiple disciplines – including biology, chemistry, business, computational sciences – need to work together and be organized into a cohesive whole. The need for fast access to a multitude of high-tech capabilities has even given rise to the “virtual biotech model”, orchestrating CROs and strategic partners.
Data exchanged via CSV files on file servers (or even in emails) is woefully inadequate for the needs of modern drug discovery – and yet that is how many biotechs still share their data internally. It’s an unnecessary barrier to collaboration that spawns inefficiencies, data-quality issues and version-control headaches.
And this is where the cloud offers an immediate qualitative advantage. Experimental data can be ingested by a cloud database with ease. With a bit of effort, this can even be fully automated. Easy-to-use dashboards, built using self-service tools (no coding), make access to that data both straightforward and controllable, from scientists to executives. Moreover, cloud-based data infrastructure is the first step towards getting ready for machine learning and AI technology – which, critically, rely on high-quality consistent data.
Advantage 3: Reduce time from experiment to insight
Hypothesis, experiment, insight – that’s the feedback loop that drives discovery. Any structural change that can make this flywheel spin faster has an exponential effect down the road.
Although hypothesis generation remains the preserve of human ingenuity, many processes around experiments and data handling are absurdly robotic. Data being shared between people or departments might need to be downloaded from one server, converted to another format and then uploaded somewhere else, for example. Inefficiencies like these, which prevent scientists having the data they need, when and where they need it, should be addressed ruthlessly by savvy biotechs – and this is something at which cloud-based systems excel.
By their nature, they are set up for connectivity with other software, via open APIs, and can host any form of data processing. They are primed for automating data pipelines and even come with convenient front-ends for data analysis.
A typical strategy is to start by storing data in a cloud, then build data ingestion connectivity (to CROs, for example) step by step.
How to get started in the cloud
It’s straightforward and relatively cheap to make those first steps, partly because any scaling up can be done incrementally, in line with your needs. But who should you trust with your data?
Amazon, Google, or Microsoft?
Since the cloud market has matured, the fight for market share has become more intense. Cloud providers such as Amazon Web Services, Google Cloud Platform and Azure (Microsoft) are falling over themselves to offer support (and free credits) to steer you away from the competition.
As with any technology that will hold your precious data, lock-in effects must be considered. How hard is it to switch providers? Moving data out of the current cloud (“egress”) is much more expensive than moving data into the new cloud. Compared with the lock-in hurdles a tech company would face, however, it’s a negligible issue – and it may never happen. It’s probably best not to overthink this scenario and cross that bridge should you ever come to it.
Of the three big players, the one you’re most likely to be half-embedded in already is Google – whether that’s just your email systems or the docs, sheets and more of Google Workspace. Continuing on this path will be the logical next step for many. User management – stuff you really don’t want to think about – is taken care of automatically, and you can expect anyone already familiar with Google sheets to be building quick frontends for data in no time.
The other thing to consider is the expertise you’ll need.
Do I need to hire dedicated cloud engineers?
In short, no. For anyone with a bit of programming experience, learning to use cloud technology is fairly straightforward. Even so, it can’t hurt to screen for cloud experience in recruiting for computational roles.
Ensuring the architectural decisions concerning the setup of your systems are future-proof, and getting automated data pipelines up and running quickly, are slightly different matters. Here, you can either work with experienced freelancers or bring in a specialized partner. Alternatively, build it yourself and re-engineer later; anything you build in the cloud is not set in stone, and it’s most important to just get going.
Whichever option catches your eye, we’re happy to share our experience and offer advice on how to best navigate this technology.
