Figure 3.

Dispersion and histogram plots visualizing sample quality and noise within the data. (A) Dispersion plot of the mean of the normalized counts was plotted using DESeq2. The plot estimates dispersion or intra-sample variability in a gene’s expression within each condition group (Control vs. UV-injured). Interestingly, analyses showed a high number of low-count features at the limit of the y-axis for estimated dispersion, and (B) the histogram of the log₂ count data vs. number of genes expressed also displayed similar low-count features across all samples, even after removal of 0-count genes from the dataset. This indicated a robust degree of sensitivity due to the high sequencing depth (see Table 1) and may also indicate the detection of low-copy-number transcripts, long noncoding RNA’s, and additional species of transcript that can be the subject of future investigation. A conservative pre-threshold limit for counts across all samples (≥20 counts for each of the six samples (n = 3 pooled samples/condition (Control vs. Experimental)) was then established (visualized by the red dotted line). (C,D) Dispersion and histogram plots following the established threshold limit for counts across all samples to eliminate noise, showing the removal of low-count features.