Fig. 1
From: JOINT for large-scale single-cell RNA-sequencing analysis via soft-clustering and parallel computing
Overview and convergence tests for the JOINT algorithm. a Workflow of the JOINT procedure. Soft-clustering, parameter optimization and DEG analysis are performed simultaneously in JOINT. Probability-based soft-clustering for cell-type identification and DEG analysis are demonstrated in the insets. b Convergence of πk (k = 1), qg,k,l (g = 0, k = 0, and l = 1), αg,k,l (g = 0, k = 1, and l = 1), and βg,k,l (g = 1, k = 0, and l = 1) to true values with iterations. c Convergence of π1, q0,0,1, α0,1,1, and β1,0,1 to true values with the number of samples. d Convergence of π1, q0,0,1, α0,1,1, and β1,0,1 to true values with dropout probabilities. True values are indicated by blue lines. Error bars in (c) (d) indicate the full range of data variation