Welcome to our first edition of Bio Endeavors! The team at Innovation Endeavors is starting this newsletter as a way to share what we’re learning with our community and instigate discussion.

Most of you know us already. We’re an early stage venture firm with a central thesis that technological advances across engineering, data, and computational approaches will converge and drive rapid transformations for our economy and society - including domains previously untouchable for venture capitalists. We call this phenomenon the Super Evolution.

Nowhere is this more exciting than the life sciences (broadly defined, as you will see). While we’ve been harnessing the power of biology for millennia, from brewing beer to breeding plants to utilizing natural medicines, the rate of advancement we’ve seen over the last several decades is entirely unprecedented. Tools built at the convergence of chemical biology, bioengineering, computation, and many other disciplines have begun to completely reshape our world, from the medicines we increasingly design to the food we eat to the physical inputs in our global economy.

We believe that the “super evolution” of biology will transform this historically bespoke field into an engineering discipline. We imagine a future in which we can efficiently design, build, test, and learn from the single molecule to systems-level ecologies. We see the fundamental advances enabling this in three categories. First, it is hard to understate the importance of computational tools to drive new biological inferences from today’s incredibly disparate set of data types, from mass spectrometry to bulk sequencing to single-cell multi-omics to clinical data and many more. Second, new experimental techniques and tools are critical for generating data in representative model systems, which in turn underlie our ability to understand living systems; more and more, our experimental design should be built to take advantage of machine learning approaches (as in this great example from de Boer et al., 2019). Finally, we will need to develop and deploy novel biological tools and useful abstraction hierarchies to more rapidly and precisely manipulate biological systems based on this knowledge. Drew Endy, who first made a case for standardized tooling and abstraction hierarchies in 2005, put it best on Twitter: “CRISPR is to synthetic biology as a cordless drill with interchangeable bits is to carpentry.”  

We have had the privilege of backing incredible scientists and companies at the forefront of this revolution - Eikon Therapeutics, Dewpoint Therapeutics, Ukko, GRO Biosciences, Character Bioscience, Zymergen, Bolt Threads, Freenome, and many more. We hope to continue to work alongside the next generation of founders passionate about these problems.

Each month, we plan to share our thinking on one topic of interest - from the intersection of climate and synthetic biology to ML-enabled therapeutic discovery to the future of cheese to living medicines to data infrastructure in the life sciences. The list goes on and on. We'd love to hear from you if you have specific topics of interest and/or are working on these problems.

June 2022 Issue: The next 10(0) years of engineering plants

While much of the focus in synthetic biology has been on work in microbial hosts, most of our agricultural production today is in plants. Plants provide the food we eat, clean the air we breathe and are a rich source of biological and chemical diversity. We believe there is a real opportunity to leverage plants and their biology to drive a more sustainable economy. Engineering plants to produce even more food, be resilient to increasingly frequent and severe climate-related events, sequester carbon, provide additional nutrition, and generate novel chemistries could make a huge difference in our ability to thrive on this planet. And yet, despite meaningful recent advances and almost a century of prior work, design-build-test-learn cycles still take too long to deliver the gains we need.

Alongside our colleagues at Leaps by Bayer, we recently brought together leading minds from across academia and industry to review recent advances in plant synthetic biology, highlight opportunities for the future, and discuss how we can mobilize resources against those.

Today, we are excited to share some of our learnings from those conversations and perspectives for the future.

Brief history → People have been developing plants to their advantage for thousands of years, primarily by breeding plants for domestication. Genetic manipulation was first demonstrated in the 1920s via chemical mutagenesis and then dramatically accelerated with the discovery of Agrobacterium-mediated transformation in the 1980s.

The 40 years since have brought tremendous progress. 90%+ of corn and soy cultivated in the U.S. is now genetically modified, but at the same time, there are only 10 crops with genetically engineered traits on the market in the U.S. today despite 40+ years of research. A study by Philips McDougal estimates that it takes 20 years and $100M+ to bring a single biotech trait to market. Further, the seed industry has consolidated dramatically, and we’ve seen public and regulatory pushback against GMOs and the consolidated chemical-intensive system of agriculture that GM crops have come to represent. In aggregate, these difficulties made the world of plant engineering a challenging field for venture investors. New technologies and potential avenues for value creation are now driving rapid change – and, we hope, an inflection point – in the field.

Fast forward to the present →  On the technical side, new tools (e.g., CRISPR editing tools, ML-based approaches to gene discovery) are already beginning to revolutionize what’s possible. However, important technical challenges and tensions remain.

A few themes that stand out to us:

• Targets remain a bottleneck. Significant basic science questions are elusive. It takes much bespoke analysis to elucidate complex pathways (not to mention ploidy and genomes 50x our human genomes), and it’s significantly harder to move beyond the genome. We need to continue to develop tools that drive understanding across multiple levels of abstraction (e.g., genome -> multi-ome -> single cells -> systems -> ecosystems). While computational approaches can provide a lot of leverage here, we need methodologies that help us learn more with less data, and we need tools (i.e. editing and regeneration) to generate more data faster.
• The rate at which we can learn about plant systems is limited by our editing and regeneration tools. We don’t yet have a floral-dip-like process in most plants, so transformation and regeneration are manual and slow processes. Perturbation feedback loops are much too long. Editing efficiently across many cell types and beyond the genome would enable more control and more questions to be asked. Developing new and faster ways to measure complex phenotypes would enable increased iteration. Ultimately, creating radically more diversity through higher-throughput editing and regeneration systems would allow us to create better computational models, helping address the target bottleneck. We are keen to understand better what a truly breakthrough system here might look like and work with anyone digging into this problem.
• Better model systems - and cell culture - will be an important piece of the puzzle. Still, challenges remain here as well (i.e., all models are poor, but some are useful): Protoplasts can be a great model for studying regulatory pathways. Still, somaclonal variation remains tough to manage, and translatability to higher plants often remains unknown. More generally, cell culture systems could serve as interesting production methods for various products and ingredients; plant metabolic pathways are complex and provide an enormous source of chemical diversity that is hard to access via other means. We are keen to understand the potential and the challenges here better.

Looking towards the future →  We think there’s lots of exciting work to be done in this space, and we will have to go beyond yield in row crops.

A few things we’ve been thinking about:

• Leveraging natural diversity is an important starting point → While the last century of plant breeding has seen incredible achievements in yield, we’ve also created homogenous populations that are increasingly challenging to optimize. Similar to work in human genetics, natural diversity with well-characterized phenotypes provides an essential and under-tapped resource for understanding and engineering relevant traits.
• Alternative food applications require innovation → Global retail sales of plant-based meat crossed $5B in 2021, and protein content is seen as a key-value driver in many key input crops. Perhaps equal opportunity is held by the potential for plants to enable new categories of alt-food. For instance, according to GFI, plant-based milk accounts for 15% of retail milk sales in the U.S., while plant-based cheese currently accounts for <1% of retail cheese sales. Numerous players are working on unlocking the category by discovering or engineering proteins within plants to serve as functional substitutes for casein (e.g. Nobell Foods, Climax Foods). Finally, we hope to see more sustainable and resilient alternative products developed in categories like coffee and chocolate. Here, underlying crops 1) drive high deforestation rates and 2) face growing supply risks in a changing climate. While alternative products may or may not be produced through engineered plants, we are confident that having a better understanding of the pathways and chemistries underlying the functional properties that we care about in these crops will be part of the solution.
• Climate and carbon are top of mind→ Climate resilience is critical. With commodity prices as high as ever and the threat of losses due to climate stress, growers are willing to experiment more in order to maintain yield in light of stressors or higher input costs. We should also be thinking harder about how to maintain output with lower inputs and how to grow on marginal lands. Carbon sequestration in agriculture holds the potential to draw down over 3 Gt CO2(e) per year by 2030. For growers, carbon offset revenue also holds an opportunity, with certification on the part of major registries being a vital piece of the puzzle (e.g. can we get registries to certify based on the change in products and practices?).
• The broader bioeconomy points to applications beyond ag; here, we are just scratching the surface → Interesting applications include leveraging plants as hosts for natural product production; more functional materials for apparel or other fibers; more efficient forms of industrial feedstocks; and highly performant, grown building materials. We want to continue to see creative approaches that leverage these fantastic foundations.
  ‍

So what? Meaningful businesses are already being built here, but long-term success will take technical breakthroughs paired with careful value chain management. The achievements and challenges of early startups in the space (e.g., Pairwise, Benson Hill, Inari, Calyxt, and more) hold essential lessons for new entrants. Entrepreneurs will need to 1) find ways to connect with end-buyers and de-risk the value chain and value capture, especially if work goes beyond yield, and/or 2) introduce a scalable set of tools that can drive a step change in test-learn cycles across crops and applications.

We’re keen to engage with anyone:

• Tackling the core technical bottlenecks, especially editing, transformation and regeneration.
• Doing pioneering work to better understand and engineer beyond the genome (e.g., (epi-)transcriptome, (epi-)proteome, metabolome).
• Leveraging plant chemistry to create new-to-world products.
• Coming up with interesting and creative ideas.