Galagan Lab

Roman’s work demonstrated cutting edge wearable sweat monitoring with a novel biosensor in a fully powered human study, paved the way for multiplexed novel analyte sensing, and introduced methods to utilize EIS as a systems identification tool for biosensors. Congrats Roman!!

🎓Congratulations to Dr. Roman Addokhi on completing his PhD dissertation titled “Development and Validation of Wearable Devices for Real-Time Sweat Biochemical Sensing.” #PhDDefense #WearableTech #Biosensors #SweatSensing #BostonUniversity

Thanks to all who contributed! @RegulonDB @wade_lab @PatDotPy @VhTierrafria

We compared BoltzNet to existing state of the art methods, and BoltzNet outperformed - demonstrating the ability of a biophysically designed neural network for interpretable machine learning and state-of-the-art accuracy. We’re proud to share this resource with the community.

Our results provide insight into several global features of TF binding behavior, including clustering of binding sites, the importance of poorly conserved accessory bases, the physiological relevance of weak binding sites, and the background affinity of the genome.

We generated binding profiles for 139 TFs—nearly half of E. coli’s repertoire. All data is public: 📂 Raw → GEO 📈 Processed → @RegulonDB 🧠 Models/tools → boltznet.bu.edu Includes model viewer, browser, and access to our evolving codebase.

🚨 Our BoltzNet paper is out @NatureComms! nature.com/articles/s4146… We are excited to provide a website with access to all models, code and data boltznet.bu.edu

We describe BoltzNet, a biophysical neural network that predicts transcription factor (TF)–DNA binding from ChIP-Seq data by directly mirroring a quantitative model of molecular interactions.

Applied to over 125 E. coli TFs, we quantitatively designed novel binding sites, and gained insight into the roles of clustered binding, bases outside of the core motif, and the affinity of the genome. We hope this tool and resource will be valuable to the scientific community.

Pleased to report our paper on #BoltzNet, a novel biophysically-designed neural network to quantitatively predict TF binding affinity to any DNA sequence. Many thanks to our collaborators @wade_lab and @RegulonDB biorxiv.org/content/10.110…

We look forward to continue sharing the exciting work our lab is doing, and thank all of our wonderful researchers and collaborators for their hard work pushing the boundaries of science across our areas of study!

And most recently, in collaboration with @Grinstaff lab led by @ukuzma2 and @ksankar93 we published a tutorial review discussing strategies for building and optimizing biosensors! pubs.acs.org/doi/abs/10.102…

Hello World! We’ve got several exciting updates to catch up on since our last posting. Time for a quick 🧵