Skip to main content
Fig. 6 | BMC Bioinformatics

Fig. 6

From: GMMchi: gene expression clustering using Gaussian mixture modeling

Fig. 6

Example output of the GMMchi python package: a log2-transformed distribution of the gene expression level of CDH1 (E-Cadherin) with its fitted normal components along with a red vertical cutoff from the tail. b BIC score. The red being the BIC score for fitting 2 normal components, the blue being the BIC score for fitting 1 normal component and the grey dashed line indicating which model to pick at each iterative step. c Bar graph showing the percentage of samples under each categorized group. d Graph showing the χ2 value calculated at each iterative step during iterative tail trimming. The red line represents the χ2 value of fitting 1 normal component, the blue line represents the χ2 value of fitting 2 normal components, and the red arrows indicate when a better fit is found. The best fit with the lowest χ2 value and the lowest BIC is with 9 low tail values removed and just one normal distribution for the remaining data

Back to article page