Additional file 8: Figure S1. of Genome-wide identification of conserved intronic non-coding sequences using a Bayesian segmentation approach Manjula Algama Edward Tasker Caitlin Williams Adam Parslow Robert Bryson-Richardson Jonathan Keith 10.6084/m9.figshare.c.3727531_D10.v1 https://springernature.figshare.com/articles/figure/Additional_file_8_Figure_S1_of_Genome-wide_identification_of_conserved_intronic_non-coding_sequences_using_a_Bayesian_segmentation_approach/4793191 Model selection for eya1. Approximations to well-known information criteria AIC, BIC and DICV for 1-12 classes. Generally, a lower value of the information criteria indicates a better model. BIC clearly suggests a 3-class model. The first local minimum of AIC and DICV has also occurred at the 3-class model. Therefore we selected a 3-class model for this data. (TIFF 48 kb) 2017-03-27 05:00:00 ncRNA Conserved non-coding sequences Putative functional elements Genome segmentation Bayesian modelling