Skip to main content
Fig. 7 | BMC Bioinformatics

Fig. 7

From: Benchmarking imputation methods for network inference using a novel method of synthetic scRNA-seq data generation

Fig. 7

Impact of imputation on network inference performance and gene-gene correlation preservation for 100 different simulated 50 gene networks using sparsity = 0.02 with multiplicative regulation for various capture efficiencies. Figures A–D show boxplots of precision scores obtained for different imputation algorithms displayed on x-axes for PIDC, CLR, GENIE3 and Empirical Bayes respectively. RAND corresponds to precision obtained using random classification and GT data corresponds to precision obtained without downsampling (i.e., capture efficiency is set to 1). Figures E–H show the mean squared deviation between gene-gene correlations obtained using the ground truth data and those obtained using various imputation methods displayed on x-axes (with results plotted on a log-scale). Figure E show results obtained using all reaction types, while Figure F, G and H show results obtained using only activation, inhibition and non-reacting type reactions respectively

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us