Oral Pre-malignant Lesions
A total transcriptomic approach to profile human and microbial transcripts in the characterization of pre-malignant lesions compared to oral squamous carcinoma (cancer/tumor) and normal (control). Built a cross-talk model to link significant human pathway activity to microbial abundance. Applied various ML approaches such as Random Forest for imputing missing smoking labels, and PCA/UMAP for dimensionality reduction.