Biotechnology is one of the latest market areas to feel the impact of software. More and more information is being digitised in the world of big pharma, with technology being used to accelerate processes and innovation, as well as reduce errors and costs. The concept of ‘infrastructure-as-code’ (IaC) from enterprise IT has crossed over. Welcome to a brave new world I call ‘biotech-as-code’.
IaC benefits have long been proven: Knowledge can be encapsulated once and then made easily repeatable, scalable, and maintainable. Therefore, results become more consistent, and dependency on risky manual effort is reduced: Goodbye to carrying out a test process nine times only for something to go wrong on the tenth occasion. Of course, ‘as code’ is heavily dependent on software automation, but increasingly, also on AI and machine learning.
For instance, AI techniques can be used to predict protein folding within DNA, something that was a grand, nigh impossible challenge that only recently has been solved. Using proven software techniques, the DNA sequence can be converted into a text string, even if it is a billion characters long, and start to analyse it more efficiently. Patterns such as repetitions, repetitions in relation to others, and when they occur can help biotech professionals learn more without needing to mix chemicals in a lab and perform years of manual tests. Suddenly, biochemistry looks a lot like software programming.
Another advancement is robotic labs, where instead of humans mixing chemicals, robotic arms carry out those tasks and report the results. Taking this a step further once a knowledge base of those chemical interactions is established. Instead of running all the possible experiments in the robotic lab, those tests can be simulated in AI environments.
The possible advantages are powerful: simulation can be used to see how millions of new drug compounds could work in the body long before testing on human subjects is approved. With most combinations already discounted, the effort can be focused on just a few most likely candidates, meaning that years have been knocked off previous processes and reduced overall cost.
Theory aside, some issues still need to be understood and addressed. Pharmaceutical firms may be recruiting software engineers, but they compete with every other market for limited talent. Biotech-as-code will mainly be performed by people who are not trained software engineers and on a colossal scale. Plus, there is the issue of keeping track of data and making it visible to others. Old-world processes such as sharing knowledge via email or spreadsheets are just not going to work.
There is also safety and security to consider: vulnerabilities that occur during code creation can be exploited by hackers for malicious purposes, could lead to malfunction, or open the door for confidential data to be illegally sold on the black market. Furthermore, biotech is heavily compliance-driven, so any software involved must be auditable.
So, why do biotech firms need to rapidly adopt software development techniques and tools in widespread use in enterprise IT? Examples include using coding standards to improve security and safety, both static and dynamic automated testing, and real-time collaboration via diagramming tools. In addition, version control helps to show who changed what, where, when, and how, while traceability management makes it easier to demonstrate the entire history of a project for regulatory purposes and internal project management.
There is also a wealth of resources – often free – available to anyone involved in software development to create better awareness of some of the biggest and most current security risks. For example, the OWASP Top Ten, is a list of the most common vulnerabilities in open source software and web applications. Another valuable resource is the National Vulnerability Database (NVD), which while managed by the US, is a valuable resource globally, and includes databases of security checklist references, security-related software flaws, and vulnerabilities. Further resources include the Open Source Security Foundation and the Security Scorecard project from the Open Source Software Foundation.
Once the biotech industry builds the same solid software best practices already available to enterprise IT, it can address accuracy, scale, and time restrictions – providing it with the foundation for far more innovation. For example, scientists can not only understand DNA but also fix defects, prevent diseases, implement reverse-aging, and improve people’s memories. Furthermore, a person’s DNA profile could be connected to a food company so that highly tailored nutritional advice can be provided. DNA analysis for individuals interested in knowing more about their genetic history is already reducing in cost and speed, and that is only going to continue.
Whether or not everyone approves of all these developments is another topic, but one thing is for certain: we are talking about massive acceleration. In terms of impact, this is comparable to quantum computing. Within a couple of decades, we will probably look back at the 2020s and consider them the Dark Ages of Biotech. Of course, most innovations will start where the money is, but ultimately, the potential benefits for everyone on the planet could be huge. Regardless of what the future may bring, the latest in software developments will play a critical role in enabling the future of biotech, bringing the benefits experienced in other industries – particularly overcoming complexity at scale – to this exciting and growing market.
By Rod Cope, CTO, Perforce Software.