Skip to main content
Fig. 3 | BMC Genomics

Fig. 3

From: The effect of methanol fixation on single-cell RNA sequencing data

Fig. 3

Differences in statistical features of genes with the top contribution in driving variation between live and fixed cells. (A) Comparison of relative expression of 500 genes with the top contribution in PC1 and PC2 between live and fixed cells. Expression of PC1 genes correlated well while in PC2 the trend was incoherent for genes with different expressions level, which indicates genes heavily loaded in PC2 may be responsible for the separation between two groups of cells. (B) Comparison of expression variation of genes with top contribution from PC1 and PC2. In the top panels of Fig. 3B, we take genes that are heavily loaded in PC1 respect for live cells and fixed cells. Then, we computed the coefficient of variation (CV) of each gene across all cells. The CVs for each gene are then plotted against that gene’s mean expression level, separately for live (blue) and fixed (orange) cells. Genes with the top contribution in PC2 holds much higher variation compared with PC1 genes. (C) Comparison of gene detection number after expression filtering. A series of thresholds were set up for different sensitivity requirements. The detection number in fixed cells gradually surpass live cells once the threshold increased (nsP > 0.05, *P < 0.05,**P < 0.01,****P < 0.0001). (D) Relative abundances of genes with low (< 5 TPM) or high (> 30 TPM) expression, the inset bar charts compare the quantities of genes which have higher expression in either live (blue) and fixed (orange) cells. For low expression genes, they are generally more abundant in live cells. Genes with higher expression are more abundant in fixed cells

Back to article page