Embracing Fungal Complexity in Mycelium R&D
What does it mean to work with the full breadth of the fungal body? To, insomuch as it is possible, work with the thallus of a fungus, the mycelium, in its full form, intact, expressing its full physical vocabulary? And beyond that, what does it mean to attempt this within human social, technical, and economic constructs that ebb and flow in their alignment with the depth and challenges of this fungal physicality? All toward the goal, ultimately, of identifying and cultivating a product, at scale, with human utility from this physicality?
Mycelium R&D is a beautiful, enriching, brutal, and disjunctive technical and cultural space. It provides endless opportunities for the driven multidisciplinarian, while simultaneously being highly sensitive to cultural, technical, and economic over-excitement and risk. It rewards breadth and patience, then punishes haste. It invites creative ambition, then exposes every hidden assumption in the system you built around it.
To work in this space is to accept that the organism holds its own logic of behavior, context, and memory, and that our success depends on how well we can interpret and negotiate with that logic. It requires a kind of technical humility, an ability to see the fungus as a living, recursive system that responds to every decision we make.
Working in mycelium product and process R&D requires blending elements of mycology, process engineering, design, systems thinking, materials science, and data science. As a result, it’s not always intuitive which skill sets or frameworks are most essential. What can we view as generalizable to practical, real-world, mycelium R&D?
I’ve come to recognize a set of organizing principles that is far from comprehensive, but that I believe can be generalized to many practical mycelium development scenarios. They are offered here as a practitioner’s interpretive framework: a way of organizing what I have found to sit near the foundation of working effectively with fungal systems.
Physical (phenotypic) plasticity is the heart of designing with filamentous fungi and defines both the critical opportunity and greatest challenge. The opportunity lies in being able to engineer diverse, multi-feature outcomes from a wide range of addressable inputs, but this same responsiveness can make processes uniquely difficult to control and scale. Fundamentally, physical plasticity derives from the nature of filamentous growth itself, where dynamic decision making drives how the organism distributes itself in space (Moore et al., 2011; Slepecky & Starmer, 2009). This behavior is a strategy that increases the probability of continued resource acquisition while maximizing efficiency. In practice, this strategy expresses through three linked features (Fricker et al., 2017; Moore et al., 2011):
Exploratory growth distributed across many independent tips, each functioning like a locally responsive decision-making node interpreting local conditions.
Continuous adjustments in branching, fusion, directionality, and mass flow, allowing redistribution of biomass toward or away from opportunity.
A guiding logic of efficiency, where spatial investment reflects the changing economics of resource availability.
Designing with fungi, then, is not about controlling a single variable or targeting a single response feature, but rather about embracing and working within a high-dimensional system, consistent with broader phenomics approaches that treat phenotype as a high-dimensional object (Houle et al., 2010). Fungal responses are inherently multidimensional and interdependent, with each phenotype reflecting the influence of multiple interacting parameters. Much of this dimensionality comes directly from fungal biology: the decentralized nature of filamentous growth, the sensitivity of hyphae to local gradients, and the nonlinear way that spatial structure, nutrition, and environment shape one another in real time. In practical R&D terms this leads to truly immense solution spaces.
To navigate this dimensionality we need a technical toolkit that can handle complexity without becoming overwhelmed by it. That means becoming comfortable with featurization, feature engineering, dimension reduction, and feature importance (Jolliffe & Cadima, 2016; Kuhn & Johnson, 2019), not just as data science principles, but as ways of translating the physical language of fungi into learnable and practically manageable signals that capture the depth of physical responsiveness.
The breadth of input-response spaces means adopting efficient and responsive experimental design, like adaptive design and learning-based approaches (Chaloner & Verdinelli, 1995) that can maximize insight and distance of actionable progress per experiment, recognizing that single factor and full factorial experiments are rarely the most efficient options. And crucially, it means learning to separate the speed of experimentation from the speed of learning. Leveraging uncertainty reduction and model confidence as signals can help prevent futile searches by showing when we are gaining understanding, and more importantly, when we are not.
Designing with and within fungal complexity is foundational, driven both by what can be mined from fungal physicality, and by a respect for how easily bad conclusions, futile searches, and failed programs can emerge in such a complex development space. Bringing together the language of phenotypic plasticity with a robust learning toolkit is about showing respect for the organism, for the practice, for the potential of the industry, and for humans.
References
Chaloner, K., & Verdinelli, I. (1995). Bayesian experimental design: A review. Statistical Science, 10(3), 273–304. https://doi.org/10.1214/ss/1177009939
Fricker, M. D., Heaton, L. L. M., Jones, N. S., & Boddy, L. (2017). The mycelium as a network. Microbiology Spectrum, 5(3), Article FUNK-0033-2017. https://doi.org/10.1128/microbiolspec.FUNK-0033-2017
Houle, D., Govindaraju, D. R., & Omholt, S. (2010). Phenomics: The next challenge. Nature Reviews Genetics, 11(12), 855–866. https://doi.org/10.1038/nrg2897
Jolliffe, I. T., & Cadima, J. (2016). Principal component analysis: A review and recent developments. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 374(2065), Article 20150202. https://doi.org/10.1098/rsta.2015.0202
Kuhn, M., & Johnson, K. (2019). Feature engineering and selection: A practical approach for predictive models. Chapman and Hall/CRC. https://doi.org/10.1201/9781315108230
Moore, D., Robson, G. D., & Trinci, A. P. J. (2011). 21st century guidebook to fungi. Cambridge University Press. https://doi.org/10.1017/CBO9780511977022
Slepecky, R. A., & Starmer, W. T. (2009). Phenotypic plasticity in fungi: A review with observations on Aureobasidium pullulans. Mycologia, 101(6), 823–832. https://doi.org/10.3852/08-197