Research Interests

The most appealing examples of complexity are, in my opinion, found in biology. Hence, although I started my Ph.D. in theoretical physics under the direction of Nigel Goldenfeld, my initial research delved into an ecological puzzle: the Species Area Relationship (SAR). This striking regularity, known for over a hundred years, states that as larger and larger areas of an ecosystem are sampled, the number of species contained in those areas grows as a power law with a coefficient close to one quarter (S~A1/4). This pattern appeared to be universal, applying across all forms of life, from bacteria and plants to mammals and birds.

We demonstrated that this universal trend is a robust consequence of species' spatial and abundance distributions. Specifically, the S~A1/4 relationship is expected when individuals of a species are highly clustered and their abundance distribution approximates a lognormal (a frequently observed phenomenon). Therefore, this trend is rather independent from the particular dynamics of the individuals in the ecosystem (competition, dispersal… etc), leading to its apparent universality. 

I was initially captivated by the complex examples of biocomplexity in macroecology but also discouraged by the difficulty of conducting real-time experiments in that setting. This led me to pivot my focus to the rapidly growing field of microbial ecology. My initial work in this area involved a close collaboration with geomicrobiologists, where we investigated the influence of microbial ecosystems on terrace formation. A central question was whether microbes were responsible for these formations and if they could serve as indicators for potential landing sites on Mars. In this project, I developed and applied new and existing methods for estimating microbial abundance and diversity using 16S rRNA gene surveys.

While the metagenomic study of this microbial community yielded a detailed blueprint of the metabolic pathways present, this knowledge was descriptive in nature, rather than predictive. Questions such as: “which species will become dominant if a given condition (e.g. pH, or acetate availability) is altered?”, or “what will be the biochemical impact of a community on its environment?” are not answerable from just the knowledge of the genomes (or even transcripts, proteins or metabolites) present in a microbial community.

We soon realized that the next step in terms of developing quantitative predictive models involved studying metabolic fluxes. Prediction of metabolic fluxes (i.e. the rate at which molecules proceed through a reaction per unit of time) for all metabolic reactions in a given organism entails knowledge of growth rates (from the carbon fluxes to biomass) and metabolite excretion (from outgoing metabolite fluxes).

To pursue this, we established a collaboration with Jay Keasling, whose group operated an EBPR bioreactor and had previously tried (with limited success due to the lack of metagenomic, metatranscriptomic or metaproteomic information) to produce a Flux Balance Analysis (FBA) model of a this system. During this collaboration, I became acquainted with an accurate technique (13C Metabolic Flux Analysis) to measure metabolic fluxes for pure cultures. This technique seemed like the perfect experimental validation tool for EBPR flux models, requiring only proper adaptation to account for the system's community nature.

I joined the Joint BioEnergy Institute (JBEI) as an LBNL staff scientist and group lead shortly after its initial funding, following my postdoc. My role focused on developing predictive quantitative models for the microbial hosts used at the institute, including E. coli, S. cerevisiae, and S. acidocaldarius. Years later, I would also join the Agile Biofoundry (ABF), fulfilling a similar role (quantitative modeling) but with a stronger emphasis on a close interaction with industry and accelerating the Design-Build-Test-Learn cycle. Developing quantitative predictive models for pure cultures seemed a good stepping stone for predicting more complex biological systems, and synthetic biology allowed us to modify the genetic code determining cell behavior.

From early on, it became clear that metabolic engineering had remained a collection of elegant demonstrations rather than a systematic discipline with generalizable methods. This limitation significantly hampered the progress of what was then considered one of the top ten emerging technologies.

Consequently, my group focused on developing general tools and methods that could be systematically applied to any host, pathway or final product. For example, we created methods to measure fluxes for comprehensive genome-scale models by using 13C labeling experiments, demonstrating that these precisely determined fluxes can enable quantitative predictions. We also showcased the practical utility of these flux measurements for better understanding production metabolic profiles and successfully making recommendations to increase biofuel production. Extending beyond single organisms, we created methods for measuring fluxes in microbial communities. For the common scenario lacking accurate mechanistic models to guide metabolic engineering, we developed techniques using -omics data combined with data mining and machine learning to increase biofuel yields. To supply the necessary large datasets for machine learning, we also contributed to the development of microfluidic chips to automate synthetic biology protocols. These techniques have been successfully applied to diverse synthetic biology challenges, ranging from the synthesis of jet fuels to the production of molecules imparting "hoppy" flavor to beer.

My current research concentrates on leveraging AI for scientific purposes, combining it with automation and robotics, and demonstrating these capabilities in synthetic biology. We work on adapting and developing AI methods to meet scientific needs: e.g., sparse data sets, feedback loops with experimental work, and combining mechanistic insight with the predictive capabilities of machine learning. We also intend to merge AI with robotics and automation to automate the creation of scientific knowledge through self-driving labs. We aim to miniaturize and automate molecular biology processes through microfluidics, and further develop and instantiate the concept of self-driving labs for synthetic biology. Furthermore, we expect to further demonstrate how AI and robotics can disruptively improve the scientific process by helping synthesize new biomaterials, fuels, and therapeutic drugs. Finally, we hope to continue to show that biology can be as predictable as physics and chemistry by tapping the right approaches.