We spent years watching protein engineers with deep domain knowledge bet on mutations that basic thermodynamic modeling would have flagged as high-risk. The wet-lab time wasn't the bottleneck. The absence of a fast, honest computational filter was. ProtSynq is that filter.
ProtSynq is a computational mutation scoring platform for protein engineers. We combine physics-based free energy methods with data-driven models trained on curated thermodynamic measurements — calibrated so the output is useful for triage, not just theoretically interesting.
The platform is built for researchers who already have structural biology intuition and just need a faster way to rank candidates before committing wet-lab hours. Not to replace experimental work. To direct it.
We're a small team. We take every support request seriously, and we maintain a direct line to the researchers using the platform. When ProtSynq gets something wrong on a structure you know well, we want to hear about it.
What ProtSynq is not: We are not a contract research organization (CRO). We do not synthesize proteins, run expression cultures, or perform wet-lab assays. We do not perform molecular dynamics simulations or FEP calculations. ProtSynq is a computational pre-screening tool — our outputs are prioritization scores, not guarantees of experimental outcome. Every shortlist we produce requires experimental validation in your hands.
PhD in Biophysics. Postdoctoral research at MIT on computational approaches to thermostability prediction and allostery, where she spent three years building per-residue ΔΔG models on ProThermDB data. Saw firsthand how enzyme engineers at adjacent labs were spending months screening variants that a simple computational filter could have ranked in hours. Co-founded ProtSynq in 2022 to close that gap.
PhD in Computational Biology. Prior to ProtSynq worked at an early-stage Cambridge computational biotech building sequence-to-function models for protein expression titer prediction. Responsible for ProtSynq's scoring infrastructure, the learned ΔΔG transformer, and the data pipeline from ProThermDB and ProtaBank curation through to ensemble calibration.
PhD in enzyme engineering and industrial biocatalysis. Wet-lab background in glycoside hydrolase thermostabilization for high-temperature saccharification processes. Joined ProtSynq in 2023 to ground the computational outputs in the practical constraints of real screening campaigns — what the kcat/Km data actually looks like after expression, and where ΔΔG predictions diverge from measured Tm shifts.
ProtSynq is based in the Kendall Square area of Cambridge, Massachusetts — two blocks from where the tools we depend on (AlphaFold2 infrastructure, protein databases, academic collaborators) are being actively developed. We're a small company in the right neighborhood.
245 First Street, Suite 1800
We respond to every message. Research inquiries, technical questions, and feedback on predictions all go to the same address.