5 Bio Predictions for 2023

Business model innovation in techbio, the year of microglia, big tech backs up the truck in bio, enzyme combos drive expansion of genome editing toolbox, and more

Welcome to Decoding TechBio, a writing collective focused on the latest scientific advancements, news, and people building at the intersection of tech x bio. If you’d like to connect or collaborate, please shoot us a note here. Happy decoding!

---

---

Happy New Year everyone! We are excited for 2023, and to share some thoughts on where bio might be headed this year. First, a caveat from a great philosopher-baseball player:

“It’s tough to make predictions, especially about the future.”

Yogi Berra

The only prediction we’re sure of is that the predictions that follow will be wrong or incomplete. Instead, we view this as an opportunity to learn and generate discussion and collaboration. Reach out with your predictions for 2023 and we’ll include them in a future newsletter. Enjoy!

1. Increased competition and a challenging financing environment will drive business model innovation in platform and AI-driven biotech

The vast majority of platform therapeutic companies pursue a two-pronged business model focused on 1) partnerships with other biopharma companies and 2) internal program development. For many years, this has made a lot of sense. Long-term value resides in wholly-owned therapeutic assets, and biopharma partnerships such as co-development deals provide short-term (upfront revenue, scientific collaboration and resources, external validation, platform development) and long-term (milestone payments, commercial royalties) value.

A couple recent trends may incentivize exploration of new business models, however. First, given the large number of biotechs launched in the last several years, the co-development partnership market is increasingly crowded, and competitive pressures may drive deal economics down. Second, with the market downturn and increasing cost of capital, biotechs may seek novel ways to differentiate themselves to raise their next round of financing.

A few alternatives exist. Platform and AI-driven biotech companies often spend considerable time and resources building infrastructure to support internal R&D and operations. Tech-forward biotechs may choose to commercialize internal their tech stack, either by licensing the platform to others, or spinning out into an independent company. In 2022 for example, Colossal Biosciences, George Church’s company with the mission of solving extinction and bringing back the Wooly mammoth, spun off an independent software company. The spinoff, Form Bio, is a platform for life sciences organizations to help manage, analyze, and visualize scientific data. 

TechBio companies may also explore the monetization of internal ML models or data assets via licensing, open sourcing, or API access. In tech, foundation models (large-scale centralized models that can be fine-tuned on more specific tasks for user-specific applications) have been on the rise. OpenAI claims to be targeting $200M in revenue next year, and $1B by 2024. OpenAI's business model and strategy is fairly secretive, but involves licensing their foundation models to other developers and companies. To be clear, replicating this strategy in bio will be challenging. Generating foundation models that are as general-purpose as models like GPT-3 will be difficult given the massive mismatch in scale of training data. Biological workflows and use-cases are often highly customized and customer-specific, so discovering a repeatable and scalable commercialization strategy for bio foundation models will take time.

There are emerging biotech business models enabled by decentralized technologies, such as IP-NFTs (harmonizing IP, underlying data, and economics such as licensing fees and royalties into one programmable, transactable, digital unit) and biotech DAOs (decentralized organizations that source, fund, and incubate translational research). Finally, while not novel business models per se, with the increasing cost of capital, we are likely to see more collaboration amongst biotechs to accomplish more with less capital. Pooling of resources, commercial partnerships or platform integrations, will help stretch the incremental dollar further. As the market retreats from platforms to products, holding company corporate structures with asset focused spinouts are likely to increase in prevalence as well.

Ultimately, the most exciting business model innovation may be something few of us expect, much less predict. We are excited to see what 2023 brings. 

2. Big tech backs up the truck in bio

How big is the bioeconomy today, and how much bigger will it be by 2030? Over the last few years, big tech has taken on the healthcare industry both organically and via acquisition with early successes in pharmacy, wearables, direct care, and cloud. We expect the usual suspects to double down in these areas. But what about bio?

As big tech places its next bets and projects where growth will come from over the next decade, we predict that leaders will focus on bio and the immense potential of the bioeconomy.

To give a few examples:

Alphabet has had an enduring effort in healthcare and life sciences across a number of departments and subsidiaries including Google Cloud, Verily, and Deepmind – it’s challenging to list them in full. At the leading edge of computational biology methods development is Alphabet subsidiary, DeepMind, acquired by then-Google in 2014, which secured its leadership in computational biology research in 2021 with AlphaFold, a 3D protein structure prediction library that helps scientists understand how proteins function from their amino acid sequences. Dating back to 2018, Deepmind began working on “reinforcement environments” for computational biology that support reproducibility in the field. We expect DeepMind spin-off Isomorphic Labs to be more active in 2023, and announce a couple splashy pharma partnerships.

Meta made waves in computational biology in November 2022 after releasing the ESM Metagenomic Atlas, an open atlas of ~620 million predicted metagenomic protein structures and an API that allows researchers to leverage the technology. This work aligns closely with Meta’s overall strategy to become a leader in AI, and the Meta AI (FAIR) group is one to follow.

So in 2023, we expect more chess moves from big tech entering and expanding into bio. DeepMind and Meta AI will continue to compete in launching new foundation models and computational tools for life sciences researchers. Alphabet, Microsoft, and Amazon will continue to duke it out for capturing life sciences cloud spend and will continue to launch new tools for analysis. And Apple? Apple will stay focused on consumer health… for now.

3. Year of the microglia

Over the past few years, there has been an explosion of papers probing the molecular underpinnings of microglia and their role in neurological disease. This body of work supports an emerging tenant in neurology: specifically regulating microglia may represent a promising therapeutic strategy across a variety of diseases with a diverse set of underlying pathogenic drivers, including aging, neurodegeneration, and inflammation. Beyond the canonical, CNS-wide microglial replacement may be an option for treating a variety of genetic brain diseases. Recent data demonstrates that a certain proportion of microglia are necessary for maintaining myelin homeostasis and integrity, which casts a shadow on the variety of CSF1R inhibitors currently being evaluated in trials. We expect to see 1) neuro-focused companies begin to build out programs focused on microglia modulation, and 2) number of platform biotech companies centered on microglia therapeutics.

4. Rapid expansion of ‘omic editing capabilities through a panoply of enzyme combinations

We will see an increased rate of publications and subsequent ventures based on fusions with gene editing enzymes. Researchers and companies are now looking for alternatives to CRISPR/Cas9 for genome editing, which are smaller in size and may have lower immunogenicity, such as CasX, Cas14 and CasΦ. But ‘effector enzyme’ discovery and development is not the only avenue for innovation in the genetic modification space. In recent years, we’ve seen an explosion of combinations of different enzymes and nucleic acids such as with PRIME editing, dCas-based epigenome editors and more recently with RADARs (10/22), PASTE (11/22), Frame Editors (BioRxiv, 12/22). In 2023, we anticipate an increased rate of new modalities arising from unique mixtures of enzymes and nucleic acids, with each breakthrough leading to its own spin-out company.

5. Accessibility of longitudinal data drives focus towards increased R&D efforts in inflammatory diseases and chronic diseases

Historically, much of biotech’s focus and investment efforts have been centered around oncology and orphan diseases, both of which often have a strong genetic basis and present acute symptoms. On the other hand, chronic diseases and many inflammatory diseases remain medical enigmas. They are often systemic, variable based on the patient and environment, and manifest as diseases of high morbidity, not mortality. Such diseases present an interesting mix of lifestyle, genetic, metabolic, and immune interactions, among other things. The bottleneck to developing treatments has, in many ways, been a lack of understanding the disease and corresponding interactions. For example, we still haven’t cracked the code on antigen presentation in the immune system—or how to match antigens to corresponding TCRS and vice versa. 

Recent advances, namely in sequencing and machine learning, are starting to make such mappings possible first by enabling the creation of longitudinal data sets and second by training models to learn the rules of the immune system and complex diseases via the data. While not a new concept, such models haven’t worked well for most multifaceted diseases because of the lack of multimodal data for training. Fortunately, this problem is starting to get solved. We predict that 2023 will see more collaborations between industry and academia to obtain longitudinal datasets for a variety of diseases such as those involving GI, microbiome, inflammation, and immune dysfunction. Government sponsored initiatives for whole genome sequencing and more open source data initiatives will spring up to help fuel discovery.

Longitudinal data and personalized treatment is at the core of precision medicine, a field that has seen many recent advances yet holds much untapped potential. We’re excited to see new developments in understanding poorly researched areas and expand the scope of precision medicine to more patients.