Large scale functional exploration of the GH5 family

The large and functionally diverse GH5 family is one of the very first Carbohydrate-Active enZYme families classified back in 1991 [3], [4]. The number of confirmed members in CAZy has grown to a staggering 27,000 where only approximately 600 (2%) have been assigned a function experimentally. In spite of early large-scale efforts [1], [2], the eye-watering paucity of characterized GH5 enzymes constitutes a general problem observed in most CAZy families, worsened by an unequal distribution of the characterized GH5 across the sequence space. This project attempts to address this problem by implementing large scale functional screening of approximately 200 carefully chosen GH5 candidates that systematically targets low or unexplored subfamilies in the known sequence space, thereby creating an excellent and large training set for use in developing progressive machine-learning algorithms that potentially can provide reliable and stable functional predictions of enzymes based on sequence information.