RGTFormer: Predicting mutation-associated multi-drug resistance in Mycobacterium tuberculosis using a categorical gated transformer and relational graph convolutional network

Key Points

• Multi-drug resistance was accurately predicted using a categorical gated transformer model, demonstrating significant potential.
• The model's predictions showed an accuracy of 92% in identifying mutations linked to resistance.
• Assessment using a relational graph convolutional network effectively captured complex mutation interactions.
• This approach highlights the need for further validation in clinical settings to enhance treatment outcomes.