Figure 2
RIP-seq data analysis. (a) Scatterplot of a control (Ctrl)-IP pair of RIP-seq data (SmB IP Lu023-Lu024), where normalized and log-transformed read numbers for each known transcript in an IP are plotted against that of Ctrl (Ctrl + 2 and IP + 2 to avoid division by zero). Black dots represent background RNAs, while the blue dots represent enriched RNAs, as determined by Gaussian mixture modeling. Only RNAs with read coverage >10 are plotted. See Figure S1 in Additional file 1 for the rest of the scatterplots. (b) Gaussian mixture modeling of the RIP-seq data (SmB IP), where the enrichment ratios for all the transcripts were plotted as a histogram (in gray) and fitted with a combination of two Gaussian curves. (c) Log-transformed enrichment ratios of the 5,296 RNAs (with coverage d >10) in all 7 experiments were clustered (average linkage clustering using correlation (uncentered) as similarity metric) and visualized as a heat map. (d) Pair-wise comparisons among all seven experiments. Numbers of enriched RNAs are listed next to the experiment labels. Black bars, number of enriched RNAs in each experiment; red bars, number of overlapped RNAs in each pair; blue bars, negative log10 transformed Fisher’s exact test P-values (within a superset of 5,296 RNAs). See Figure S2 in Additional file 1 for pairwise comparisons excluding non-coding RNAs.