Supplement Number 1

The supplementary material contains (i) values of hyper-parameters used in the XGBoost Models, (ii) confusion matrices of the various trained XGBoost Models, (iii) additional snapshots of the critical gas hydrate nucleus, and (iv) cluster plot for training data used in the O-O XGBoost Model with 3 OPs and r_cut = 5.5A, (v) Solubility of methane in aqueous phase.