Software Inspired Biological Design

...Instead of the other way around

Thinking about and doing biology is very cool. As it turns out, it’s also hard, and because hard things are typically coupled with long timelines, the end result of doing biology is that people that succeed in this lonely venture tend to go very deep in their own field (for me it was mosquito engineering, and later AAV engineering), without room for insight from others. 

Software is quite different from biology in this way. I have been personally tempted by the short feedback loops and collaborative hacker culture that the world of bits provides, and get jealous of my pure tech friends engaging with cool stuff through hackathons and weekend meetups. The end result of these dweeby gatherings, like the recent Mistral AI Hackathon in SF, is a string of exciting projects that I can only describe as pure magic. It’s unreal to think that some folks showed up on a Friday, hacked away with other engineers, and two days later, we got a completely open-sourced AI wearable that acts as an always-on digital record of every conversation you had in a day (shoutout to Nik Shevchenko). The history of hacking goes beyond software. Among many other exciting things, the first Apple computer was built in a garage.

Why can’t we, as biologists, also do this?

Opportunity for cultural change

I’ll state this upfront, because I know my lab friends will yell at me if I don’t – biology is expensive, and it takes a long time to do well. This is true! But if we flip the framing a little, what we actually get is this: biologists focus on hard problems with long timelines. I don’t mean to argue that a weekend hackathon can replicate what oftentimes to takes years or decades to accomplish in biology, but I posit that there are problems that can be tackled that lie somewhere between easy and pointless, and hard and world changing – it just requires the right people and environment!

Most scientists today get their start in academia, and whether they end up sticking around universities or transitioning to industry, they ultimately adopt a certain mental model of how to think about biological problems. It looks something like, “imagine a new way of killing harmful bacteria in people,” which often transitions to “it’s been two years and I haven’t gotten anywhere, I should consider something else in the field of microbial gene shuffling and enzymatic engineering,” not realizing just how deep they’ve gotten. Biologists need more opportunities to just step back and chat about problems in the broader field that don’t require 20 year commitments to the current paradigm. I think hackathons for biology would be a great way to start to chip at the stultified bio-culture by creating a space for biohackers to come together and think about progress in chunks. 

Mapping things out for a biology hackathon

Biology is becoming an increasingly diversified, interdisciplinary field. A single team can employ software engineers, lab scientists, and ‘omics people, which actually makes doing a hackathon pretty tractable! The key comes down to framing the right problem. Common approaches to hackathons, especially when they’re sponsored by a single big company, is that they invite participants to find creative ways of deploying existing technology. Most critical to the structure of most tech hackathons is that they are often successful despite not resulting in a tangible product. Rather, they are mostly a funded opportunity for a group of motivated people to demonstrate potential capability, and allow market forces to derive value from successful demos that result from a weekend of focused effort. To me, that is what success looks like for a biohackathon. With that framing in mind, I’m particularly interested these days in the application of Nanopore sequencers to different things, so I’ll frame most of the following around what it might look like to organize a hackathon exploring that.

I’ll emphasize here an important point: the ideal hackathon, to me, results in a thing being built, and since we’re talking about biology, this means I have a preference for atoms, not bits. Biology software is cool, but that feels more like a flavor of existing (bits) hackathons rather than a new kind of thing, so I invite you to join me in the (atoms) world for the remainder of this piece.

iGEM provides a serious outlet for high school and undergraduate students to reshape their thinking and engage in these hacker-type projects, but it feels like there’s a missed opportunity to invite heavily experienced scientists and engineers to do this type of thing as well. 

Sample bio-hackathon project themes

1. Biosecurity

   1. Radically Cheaper Pervasive Environmental Sequencing

      1. Explore optimal strategies for detecting stealth pandemics: wastewater sampling, human samples, or airport bathroom sampling?

      2. Encourage participants to engage in Fermi estimations of viral shedding and brainstorm innovative approaches.

   2. Wearable Technology and Pathogen Detection

      1. Attach Oura rings, Apple Watches, or Whoop bands to 10 participants.

      2. Investigate the potential for reliably detecting sex, age, or illness (e.g., using a harmless common cold coronavirus) based on wearable data before symptoms appear.

   3. Biosafe Food Production in the Event of an Agricultural Catastrophe

      1. Develop strategies for producing food in a biosafe environment following a bio-catastrophe that wipes out agricultural crops.

2. Affordable and Accessible Bio Tools

   1. Open-Source HiFi Polymerase or Common Enzymes

      1. Develop a polymerase that can be easily generated using common E. coli, making it more accessible to researchers.

   2. Low-Cost, Open-Source Lab Equipment

      1. Design affordable, open-source versions of essential lab equipment (think PCR machines or gel electrophoresis units) to make lab work more accessible and cost-effective for everyone.

3. Consumer Biotech

   1. Biotech Products for Entertainment

      1. There are currently 0 (!)(!)(!) genuinely commercially-facing bio products. It would be great to fix this! An idea would be to explore consumer-facing biotech applications beyond GloFish and pet cloning, focusing on affordable, non-cutting-edge technologies.

      2. Brainstorm ideas for entertaining biotech products, like glow-in-the-dark organisms or creatures that change colors based on environmental conditions, drawing inspiration from projects like glowing plants from Light Bio.

4. Combinatorial New Tech Bingo

   1. Lab Robotics, Notebooks, and Troubleshooting with Language Models

      1. Provide participants with access to large language models (LLMs) and explore their potential for enabling lab robotics, maintaining lab notebooks, and assisting with troubleshooting.

   2. Nanopore Technology Applications

      1. Give participants access to nanopore technology and encourage them to investigate novel applications and uses for this cutting-edge tool.

Scale and Location

I don’t know what a good number is for these things, but let’s say 50 people is good. Ideally they would span lab scientists, software/ML engineers, and hardware people. Unfortunately biology tends to require specialized equipment, so they would show up to an existing lab space that has flexible and open user areas. As an example, New York’s Genspace would be a great place to be, or Boston’s BosLab DIYBio space. I really don’t know what it would cost to rent out a space like this, but current monthly memberships are reasonable on the order of $50-100/mo, so I’ll imagine a slight premium and say this might cost $200/person for a weekend x 50 people = $10,000 for the space. Not bad, so far!

Current total: $10,000

Equipment

This is where things deviate a little bit from traditional tech hackathons. While specialized biomarker spaces will include commonplace equipment like pipettes, hoods, PCR machines and lab benches, most protocols making use of these things require consumables that are not cheap. As an example, a box of pipette tips (96 tips, which I’ll approximately equate to ~1 day’s spend for a low throughput experiment) can cost about $30. Enzymes, like polymerases, can cost a few hundred dollars for 1 vial. Cell culture media isn’t any cheaper, costing somewhere around $100 for 500mL depending on what type of media you need. 

What I’m getting at is these things are expensive, and careful thought is needed to not be wasteful. Additionally, most things that might require reagents also take a long time, but there are luckily a class of reagents that are cheap, like nucleotides. PCR primers are quite cheap, and while primers alone are not really anything useful, I’m hopeful that smarter people than me can imagine creative ways of working on cheaper, tractable problems that don’t require as much equipment and time as complicated cell culture / viral production experiments. But for the sake of completeness, I’ll assume that the spend of the average lab participant for a Nanopore-focused biohackathon is something like $250/day. If only 50% of the 50 people are actually doing lab experiments, and are being diligent about their spend and work for 2 days, this is $12,500. Adding another $1000 in compute across the group, and materials for anyone who wants to do more hardware focused things, we’ll bring this to $13,500 for experimental materials.

Current total: $10,000 + $13,500 = $23,500

Incentivizing success

I’m a firm believer that good incentives drive good outcomes. While I’d love to believe that all biologists are just doing this for the love of the game, I’ll also admit that being paid your worth for exceptional effort feels good. It feels important to attribute a reasonably high cash prize offering for people to work on problems, so I’ll say $5,000 cash to the winning team, $1000 to the runner up. If you’re a grad student at MIT or Harvard, a few thousand dollars for a weekend of work is really not that bad!

Current total: $10,000 + $13,500 + $6,000 = $29,500

Other logistics 

Because of the nature of biology (bacteria and cells take time to grow, enzymes take time to run) we are slightly more time constrained than our pure tech counterparts. For this reason, participants should get a budget allocation in advance, plan experiments, and order things a month or so before the scheduled hackathon weekend, so that things arrive in time to get cracking at the start. It would probably be good for people to pre-register as a team, but also some type of skill matching can occur with registered participants so that teams are a mix of talents if you want to sign up for a biohackathon as a free agent. 

Also food and stuff – I’ve only been to a single hackathon (hardware), and pizza was provided for like 2000 people for a weekend. I think we can swing some bulk purchased food for people in advance so efforts remain concentrated. For 50 people, it feels reasonable to spend $2000 for a weekend’s worth of food. I don’t have experience catering like this, so if you think it could be cheaper, you know better than me! 

Final tally: $31,500.

Final thoughts

Biology costs a lot, I know, but when you really think about it, $31,500 for high impact work actually seems very reasonable. The outcome of the latest Mistral Hackathon yielded genuine products that can enter the market, as well as tools like iterative automated prompt builders that any and all LLM users would benefit from. The material of a focused biohackathon could yield outsized returns for a field that is generally measured in decades. I would love to see the culture of biology change to invite more spontaneity and ingenuity, because I know it exists within so many stellar scientists that I’ve worked with. If you read this and find yourself thinking, “you are naive, biology just isn’t suited for this,” I invite you to consider a framework shift in your head to specific problems you solve in your own work regularly. Even simple things like primer optimization tools or true end-to-end cloning simulators would be great things of high impact! 

If you read this and are interested in chatting more, shoot me an email or twitter DM – I’d love to chat more.

Special thank you to Stephen Malina, Alex Telford, and Willy C for their helpful comments.