ForetoData / publications
Research evidence for better experiments.
Represent → Simulate → Decide → Learn
These publications test the scientific foundations of the larger decision framework: represent incomplete evidence, model and compare possibilities, decide what to test, and use results to improve the next experiment.
Functional discovery and nitrogen fixation
Evidence for finding biological mechanisms when function is uncertain, signals are spread across data sources, and experiments are limited.
Nitrosomes: protein language modeling and live-cell imaging reveal condensate-like nitrogenase organization in heterocysts
Nitrogen-Responsive Extracellular Proteomics Reveals Evidence for a Novel Heterocyst-Specific Protein Secretion Pathway in Anabaena
Harnessing Nitrogen Fixing Plants for a Bio-Solar Nitrogen Economy
Beyond nif: protein-family modeling reveals the accessory systems of cyanobacterial diazotrophy
Secondary Metabolites Predict Diazotrophic Cyanobacteria: A Model-Based Cheminformatic Approach
Discovery of Photosynthetic Oxic N₂-Fixation in Cyanobacteria Using Wet Lab and Machine Learning Approaches
Harnessing Solar-Powered Oxic N₂-fixing Cyanobacteria for the BioNitrogen Economy
Unicellular Cyanobacteria Exhibit Light-Driven, Oxygen-Tolerant, Constitutive Nitrogenase Activity Under Continuous Illumination
Identification of Cell Surface Sugars in N₂-Fixing Cyanobacterium Cyanothece ATCC 51142 Using Fluorescein Labeled Lectins
Protein learning and engineering
Evidence for using learned protein representations to compare a large search space and choose more informative variants to test.
Active Learning on Protein Language Model Embeddings Accelerates Rubisco Variant Discovery for Desired Traits
Comparative genomics and broader computational biology
Evidence for combining biological measurements across organisms, molecular layers, and collaborators to answer questions no single dataset can resolve.