Table 3.
Study | nb of genes identified in the studies | nb of genes in our filtered list | Sensitivity | Specificity | p-value |
---|---|---|---|---|---|
Ghildiyal et al. S2 | 620 | 274 | 0.44 | 0.78 | 7.6x10−12 |
Ghildiyal et al. WT head | 646 | 331 | 0.51 | 0.79 | 8.4x10−12 |
Czech et al. S2 | 95 | 51 | 0.54 | 0.76 | 2.2x10−10 |
Czech et al. ovaries | 53 | 34 | 0.64 | 0.76 | 9.1x10−10 |
Total | 1252 | 585 | 0.47 | 0.81 | 1.8x10−11 |
The numbers of genes corresponding to sequenced endo-siRNAs isolated by pull-down by Czech et al. [23] and Ghildiyal et al. [22] were retrieved according to the procedure described in the ‘Materials and Methods’ section. Briefly, all the reads in Ghildiyal et al.’s libraries were considered, and 50 reads for each sequence (as a cut-off) were retained for gene identification from Czech et al.’s libraries. The predictions are the numbers of these genes that are present in our calculated list of 4318 genes retaining only those showing the potential to form duplex RNAs with a minimum of 5 different other RNAs. The table displays the sensitivity values, corresponding to the number of computationally predicted genes found in the pull-down assays, along with the specificity values, corresponding to the number of genes that were not predicted to form a dsRNA and were not found in the pull-down assays. The p-values represent the probability of obtaining these results by chance.