Tuberculosis, caused by Mycobacterium tuberculosis, is the leading cause of infectious disease death worldwide, and is a global health burden due to the rise in antibiotic resistance. Streptomycin, an aminoglycoside antibiotic, was the first drug to treat tuberculosis. It’s long term usage has lead to the developpment of resistant strains, which have previously been attributed to mutations found in the rpsL and rrs genes encoding ribosomal protein S12 and 16S rRNA respectively. High level resistance was conferred through such mutations, however 30% of clinical cases did not present with mutations in these two genes. Low-level streptomycin resistance was recently linked to changes in gidB, a gene that encodes a 7-methylguanosine (m7G) methyltransferase specific for the 16S rRNA. The mechanism of gidB mediated streptomycin resistance remains poorly studied. Therefore, we implemented a structural approach to understand the molecular and functional consequences of drug resistant mutations in gidB. Manual data curation and genomic analysis of variants in circulating Thailand strains identified approximately 37 resistance associated and 44 non-resistance associated variants in gidB, with strong genomic and experimental evidence. The structure of the gidB protein was homology modelled, along with the 16S rRNA and drug docked to the active site. Mapping of the variants onto the protein structure revealed that they were distributed throughout the entire protein structure. The structural and functional effects of these mutations were assessed using our graph-based signature pipeline. We observed resistance associated mutations led to protein destabilization and altered binding to drug target as well as RNA, its primary substrate. These insights were used to build an empirical model to determine the susceptibility profile of novel drug resistant mutations. The predictive model was validated against an independent non-redundant test set, and demonstrated an accuracy of 80% and AUC of 0.84. We are now looking to validate our model on different clinical datasets to test the robustness of the model. In the future this tool can be incorporated in a clinical setting aiming to assist the diagnosis and management of streptomycin resistance in tuberculosis.