Functional Protein Observatory
Building big data for predicting protein function from sequence
Our Mission
Proteins—the molecular machines of life—are chains of amino acid “letters” that fold into complex 3-D structures to drive nearly every biological process. Just as ChatGPT has revolutionized text interpretation and generation, AI algorithms trained on protein structure data, like AlphaFold, now predict protein structure from an amino acid sequence. The next grand challenge in life science is learning to design proteins with desired functions—a feat that will profoundly transform medicine, engineering, and sustainability, from accelerating therapeutic discoveries to creating designer enzymes for battery recycling and biodegradable plastics. Despite recent progress in protein design, linking sequence to function is largely impossible because the required data do not yet exist.
The next grand challenge in life science is learning to design proteins with desired functions—a feat that will profoundly transform medicine, engineering, and sustainability.
To fill this data gap, Sarafan ChEM-H is launching the Functional Protein Observatory, a new initiative led by Institute Scholar Polly Fordyce to generate high-quality, multidimensional measurements of protein function. Fordyce has developed a suite of technologies for high-throughput expression, purification, and deep quantitative functional characterization of proteins at unprecedented scale—millions of measurements in days instead of decades.
Our progress
million measurements
The Functional Protein Observatory will leverage these cutting-edge technologies, to enable quantitative, megascale protein function measurements that will be deposited into an open-source database for the scientific community.
Work with us
Data only catalyzes true scientific impact when it is standardized, easily accessible, and widely shared. We aim to be a community resource for data generation for our open-source data repository. We are interested in working with internal and external collaborators from both academia and industry, who want to study a protein of interest, using the high-throughput technologies housed at Stanford. We also aim to enable other laboratories to independently generate their own data, using our bead-based platform. If you would like to discuss a potential collaboration for data generation, please contact us:
Contact us
Our leadership
Polly Fordyce is an associate professor of bioengineering and of genetics at Stanford University and an institute scholar at Sarafan ChEM-H. She is a physicist turned biologist turned engineer whose cross-disciplinary training has enabled her to engineer tools that leverage the power of microfluidics to miniaturize and parallelize experiments traditionally done in beakers and flasks. Her laboratory has developed technologies capable of generating a variety of ground-truth kinetic and thermodynamic physical constants for protein folding, binding, and catalysis at unprecedented scale. In a recent paper, her team reported more protein inhibition data in a single paper than had been reported by all researchers worldwide over the prior two years, transforming the landscape of what is possible for functional protein measurement.
Our technologies
The central hub of the Functional Protein Observatory is at Stanford University. Our technologies are capable of generating kinetic and thermodynamic physical constants for protein folding, binding, and catalysis, and we are continuing to evolve these platforms to make even faster many-on-many library measurements. The Functional Protein Observatory will house the following technologies:
Crowdsourced, distributed data acquisition using beads
To enable crowdsourced generation of quantitative thermodynamic data at global scale, we recently developed APB-Display (Amplicon/Protein Bead-Display), a bead-based platform that covalently links many copies of a DNA template with the protein it encodes for >100,000 protein variants in vitro. APBs can be generated and sorted at rates >1 kHz using common cell sorters, enabling low-cost expression and binding screens at an unprecedented scale. Analogous to the way in which next-generation sequencing transformed our knowledge of the nucleic acid sequence universe, we anticipate the distribution of APBs by the Functional Protein Observatory will enable democratized and crowdsourced acquisition of protein functional data.
Further reading
Low-cost expression, purifying, and measurement of binding affinities (Kd) for >100,000 protein/protein interactions in 3 days using only very small amounts of reagents:
- Passow, D.R., Gupta, A., Thompson, S., Kundaje, A., & Fordyce, P.M., “Amplicon/Protein Bead Display enables quantitative in vitro biochemistry at scale”, bioRxiv (2026). Link
In-house measurement of molecular binding and enzyme catalysis using valved microfluidic devices
Traditional biochemistry requires a week of labor and liters of reagents to express and purify even a handful of proteins. To scale and speed this process, we have developed valved microfluidic devices with >1,800 x 1 nL valved reaction chambers, each of which serves as a ‘nanoreactor’ for highly parallelized cell-free protein expression, purification, and characterization. Each device measures tens of kinetic and thermodynamic constants for >1,800 mutants in days using only microliters of reagents (a 1-million-fold reduction). To date, we have applied these devices to quantify different aspects of function for a variety of different protein systems.
Further reading
Quantifying kinetic and thermodynamic constants (kcat, KM, kcat/KM, Ki) for 1000s of enzyme variants interacting with many different substrates and inhibitors to understand how sequence encodes catalytic efficiency and specificity (HT-MEK):
- Markin, C.J.*, Mokhtari, D.A.*, Sunden, F., Appel, M.J., Akiva, E., Longwell, S.A., Sabatti, C., Herschlag, D.*, & Fordyce, P.M.*. “Revealing enzyme functional architecture via high-throughput microfluidic enzyme kinetics”, Science (2021). Link
- Markin, C.J., Mokhtari, D.A., Du, S., Doukov, T., Sunden, F., Fordyce, P.M. ‡, & Herschlag, D., “High-throughput enzymology reveals mutations throughout a phosphatase that decouple catalysis and transition state analog affinity”, PNAS (2023). Link
- Lee, A.A., Mokhtari, D.A., Egan, E.D., Blacklow, S.C., Herschlag, D.‡, & Fordyce, P.M.‡, “High-throughput biochemical phenotyping of SHP2 variants reveals the molecular basis of diseases and allosteric drug inhibition”, bioRxiv (2026). Link
Quantifying kinetic and thermodynamic constants (Kd, koff) for 1000s of transcription factor variants interacting with 10s of different DNA sequences to understand how sequence encodes affinity and specificity (STAMMP, k-STAMMP).
- Aditham, A.K., Markin, C.J., Mokhtari, D.A., DelRosso, N.V., & Fordyce, P.M. “High-throughput binding affinity measurements for mutations spanning a transcription factor-DNA interface reveal affinity and specificity determinants”, Cell Systems (2021). Link
- Horton, C.A., Alexandari, A.M., Hayes, M.G.B., Marklund, E., Schaepe, J.M., Aditham, A.K., Shah, N., Shrikumar, A., Afek, A., Greenleaf, W.J., Gordan, R., Zeitlinger, J., Kundaje, A., & Fordyce, P.M., “Short tandem repeats bind transcription factors to tune eukaryotic gene expression”, Science (2023). Link
- Hastings, R.*, Aditham, A.*, DelRosso, N., Suzuki, P., & Fordyce, P.M., “High-throughput thermodynamic and kinetic measurements of transcription factor/DNA mutations reveal how conformational heterogeneity can shape motif selectivity”, Nature Communications (2025). Link
Quantifying binding affinities (Kd) for 1000s of protein/protein interactions to understand how transcription factors recognize co-activator proteins to regulate transcription:
- DelRosso, N.*, Suzuki, P.H.*, Griffith, D., Lotthammer, J.M., Novak, B., Kocalar, S., Sheth, M.U., Holehouse, A.S., Bintu, L.‡, Fordyce, P.M.‡, “High-throughput affinity measurements of direct interactions between activation domains and co-activators”, bioRXiv (2024); accepted at Nature Chemical Biology. Link
- Perez, C.P.*, DelRosso, N.*,‡, Nolan, C.L., Parekh, U., Choe, C.A., Eguchi, R.R., Wen, Q., Fordyce, P.M.‡, & Huang, P.S.‡, “ADAPT-M: A workflow for rapid, quantitative in vitro measurements of enriched protein libraries”, Nature Communications (2026). Link
Quantifying kinetic stabilities (k-unfold) for 1000s of enzymes to understand sequence determinants of kinetic stability (essential for learning to predict protein half lives in the body):
- Atsavapranee, B.*, Sunden, F.*, Herschlag, D., & Fordyce, P.M., “Quantifying protein unfolding kinetics with a high-throughput microfluidic platform”, Cell Systems (in press). Link
Contact us
If you have questions or would like to work with us to functionally characterize a set of protein variants using one of our technologies please contact us to discuss a potential collaboration.