EcoFoldDB v2.0: Database of protein structures of ecologically relevant microbial functions

EcoFoldDB is a database of protein structures encoded by microbial genes involved in:Trace gas oxidation, carbon cycling (i.e., C fixation, metabolism and degradation), nitrogen cycling, sulphur cycling, phosphorus cycling, iron cycling, plant-microbe interactions, and osmotic stress tolerance. The database is used by EcoFoldDB-annotate, an annotation pipeline that leverages the scalability of the ProstT5 protein language model and Foldseek to allow structure-based functional annotations against EcoFoldDB at the metagenome-scale without needing to perform protein structure predictions.The database and pipeline are available via https://github.com/timghaly/EcoFoldDB