Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: Model selection and robust inference of mutational signatures using Negative Binomial non-negative matrix factorization

Fig. 4

Results for Po-NMF and NBN-NMF applied to a data set with 21 breast cancer patients. a The optimal number of signatures estimated from SigMoS and BIC when using Po-NMF and NBN-NMF. b The residual plots for Po-NMF and \(\hbox{NB}_\text{N}\)-NMF when assuming the estimated number of signatures from SigMoS i.e. 3 signatures in both cases. The lines in the top plot correspond to two times the expected variance under the chosen distributional assumption. As the \(\hbox{NB}_\text{N}\)-NMF holds 21 different expected variances, we have chosen to plot the median, minimum and maximum variance among the 21. The second plots show the normalized residuals. The vertical blue and red lines depict the theoretical quantiles and the gray lines show the observed quantiles

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us