> The field of biology is turning into a software discipline Advancements in technology now allow us to digitize and analyze the complex processes within the human body affordably and efficiently. For just a few hundred dollars, we can **sequence the entire genome**, revealing our genetic blueprint. Omics technologies, including proteomics, epigenomics, and glycomics, enable us to delve into proteins, epigenetic modifications, and sugar chains, respectively, providing a comprehensive view of cellular functions. Additionally, innovative imaging tools offer a closer look inside cells and help **identify crucial biomarkers** in fluids like blood, urine, and breath. With this wealth of digitized data, we can create detailed computer models of the human body. These models facilitate 3D reconstructions, simulations, and visualizations, allowing us to explore various health scenarios and interventions in a virtual environment, akin to interacting with ChatGPT. ![[Pasted image 20240316154253.png]] There is a new generation of biotech startups that are AI and Software driven such as [Recursion](https://www.recursion.com/), [Insilico Medicine](https://insilico.com/), [Generate](https://generatebiomedicines.com/), [Deep Apple](https://deepappletx.com/), [Atomic AI](https://atomic.ai/), and others. Additionally, there is a new generation of AI tools like [Betteromics](https://betteromics.com/), [Benchling](https://www.benchling.com/), [Monomer](https://www.monomerbio.com/), [Deepcell](https://deepcell.com/), [Waypoint](https://www.waypointbio.com/), … that are used to accelerate innovation in biotech and accrue value at the speed of software. ![[Pasted image 20240316152520.png]] #### 1 - Capture **Capture (read)** all interactions in biology, like analytics in internet businesses, generate massive amounts of data that give us insight into the next steps. This is crucial if we want to recreate and simulate with Software and AI more accurate digital models of living systems. This is a fast-moving field with well-established techbio hardware players that are developing faster, cheaper and better solutions such as [Illumina](https://www.illumina.com/), [Oxford Nanopore](https://nanoporetech.com/), [10xGenomics](https://www.10xgenomics.com/), [Ultima](https://www.ultimagenomics.com/), [Pacbio](https://www.pacb.com/), [Danaher](https://www.danaher.com/), [Agilent](https://www.agilent.com/), [Thermofisher](https://corporate.thermofisher.com/), …. > These companies race towards the $100 to $200 price point for whole genome sequencing. Ubiquitous sequencing will do for TechBio what Amazon Web Services did for Software-as-a-Service. #### 2 - Write **Not only can we read biological data, but we can also write new biological molecules using synthetic biology or CRISPR**: This means that we no longer need to laboriously find and combine molecules or proteins that exist in nature. Instead, now we could simply design and “print” them. > *Biology is evolving from a “discovery” based discipline, to an “engineering” or “design” discipline allowing us to use new “CAD” tools and traditional* software development cycles of design, build, test, learn, and iterate, can be applied. As the team at [Absci](https://www.absci.com/) puts it, **we are moving from finding a needle in a haystack to designing the needle.** E.g. [Humanegenomics](https://www.humanegenomics.com/) is designing with software (with their internally designed CAD tool) and “printing” oncolytic viruses — the founding team comes from the 3D printing industry. Some leaders have emerged in this category in the last few years such as [Twist Bio](https://www.twistbioscience.com/) or [Ginkgo](https://www.ginkgobioworks.com/). ### 3 - Generate & Simulate **Simulate on computers (in-silico) as opposed to in animals (in vivo) or in the lab (in vitro)** for predicting the behavior of new compounds and designs. These simulations enable us to prioritize experiments that are more likely to succeed, saving time and resources.  **Traditional bio experiments are expensive to run and take time. In contrast, with in-silico simulations, the cost ($ and time) is a small fraction**. Moreover, we can run hundreds of different scenarios in parallel, **without the need for hundreds of mice. And we can eliminate the waiting time for cells to grow in the lab or the months it takes to grow a tumor in a mouse** and observe its response to a new compound. Additionally, once the simulations are later run in vitro or in vivo, their results serve as training data for future simulations. Digital models recreating bio and molecular interactions and getting more and more sophisticated, incorporating thousands of variables and 3D visualization techniques borrowed from video games (eg. Unity) to reconstruct what is going on. One can build a simplified model of a mouse, in the same way that a mouse is a simplified model of a human. This does not guarantee that what works in the model, works in the mouse or human (living systems are complex and messy), but it assists in providing direction and **prioritizing what to test in vivo** when dealing with a humongous search space for designing new compounds, saving time and resources. Also, **to generate promising drug candidates to test, one can prompt an LLM to generate the genomic sequence** (which is a text output) with specific properties. Several startups, such as [Tessel](https://tessel.bio/) or [Absci](https://www.absci.com/), are actively working on this. > As government dollars flow in, biomanufacturing will prove to be fertile ground for public-private partnerships. [President Biden’s $2B Biotechnology and Biomanufacturing Initiative](https://www.whitehouse.gov/briefing-room/statements-releases/2022/09/14/fact-sheet-the-united-states-announces-new-investments-and-resources-to-advance-president-bidens-national-biotechnology-and-biomanufacturing-initiative/) was just the start. Given the strategic importance and capital intensity of this sector, we may see the development of Biomanufacturing Primes analogous to Defence Primes. ### 4 - Wetlabs **Cloud and automated labs (similar to cloud computing)** — In the past, starting a biotech company had a prohibitive cost for the amount of hardware equipment to buy upfront (similar to servers in data centers in the early days of the internet) and the human experts to hire to maintain that equipment. However, this is about to change. Soon, a small team in India will be able to use the best (remote) wetlabs in Boston or San Francisco and pay “by the hour”. These **wetlabs will be accessible remotely via a UI console, just as internet companies utilize AWS infrastructure, and they will be operated by robots with very little human intervention.**  The liquid handlers, mass spectrometers, bioreactors or genomics workflows now are part of the same “system” and all data produced is automatically recorded. Since the labs can be programmed and experiments executed end to end without human intervention, there are **software tools to design, schedule, run, and analyze experiments, overnight or weekends, maximizing equipment utilization, and speeding up all wetlab experiments.**  The cost of starting a biotech company is decreasing dramatically and we are about to see founders in their 20s starting companies in a garage (like internet startups). A bunch of lab automation startups, including [Monomer](https://www.monomerbio.com/), [Automata](https://automata.tech/), [Oribiotech](https://oribiotech.com/), [Limula](https://limula.ch/), [Radix](https://www.radix.bio/), and [Multiply Labs](https://multiplylabs.com/), etc, are turning this vision into a reality. ### 5 - APIfication **APIfication of the industry** to share data between cloud labs, devices, biobanks, and distributed teams, also allowing different companies to specialize in one step of the stack of operating a biotech company, so companies do not need to build the full stack (similar to what has happened in banking as-a-service). ### 6 - Biology OS **Data platforms can be combined with project management tools** (eg [Betteromics](https://betteromics.com/), [Latch](https://latch.bio/), or [Benchling](https://www.benchling.com/)) to facilitate collaboration, allowing researchers to review data, modify experiments, and build on each other’s work. [Dotmatics](https://www.dotmatics.com/?trackingcreative=606351084719&keyword=dotmatics&matchtype=e&network=g&device=c) is rolling up alot of these into solutions. ### 7 - Bio Data as a Service **Bio Data as a service** is also becoming popular, with companies like [Galatea](https://www.galatea.bio/), [Flatiron](https://flatiron.com/), [Cure51](https://www.cure51.com/), [IMU Biosciences](https://www.imubiosciences.com/), and [Culmination](https://www.culmination.com/) creating biobanks that provide access to valuable biological (multi omics) data. This allows other companies to **accelerate innovation without having to recreate these datasets from scratch**. ### 8 - Workflow Digitalisation **Digitization of various steps and workflows from filing to IND, to drug commercialization** to speed up those processes at a fraction of the cost — Examples include platforms designed for orchestrating and **digitizing clinical trials** (e.g. trial design, patient recruitment and onboarding, result collection, and data preparation for FDA submission, with companies like [Paradigm](https://www.paradigm.inc/) or [LindusHealth](https://www.lindushealth.com/)), **digital marketplaces for finding and negotiating with CROs** capable of supporting clinical trials or experiments (with companies like [Cromatic](https://www.cromatic.bio/)), and even **software platforms for market access** (i.e. facilitating price negotiations between pharma companies and payers, and managing these agreements in the future, with companies like [CCX](https://www.ccx.tech/)). [Lindus](https://www.lindushealth.com/) claims that certain clinical trials, which used to take 20 months, can now be accelerated to 12 months. Additionally, [Cromatic](https://www.cromatic.bio/) enables you to match with vetted CROs within 24 hours, as opposed to weeks. Software in biotech allows you to do more, with less and faster. Watch for the new data standard that the industry coalition [Transcelerate](https://www.transceleratebiopharmainc.com/initiatives/clinical-data-standards/) is putting together. ---- Links: [[Biologics]], [[Biomarkers x Biological Age]] [[Decentralized Science]]