Table 3. Logistic regression of P-element copy number and structural variation with hybrid dysgenesis after accounting for differences in hobo-abundance.
| P-element Estimate | hobo Estimate | P-Value (Full Dataset) | Association (Full Dataset) | P-Value (No DGRP-42) | Association (No DGRP-42) |
|---|---|---|---|---|---|
| Normalized coverage | Normalized coverage | 2.73E−01 | − | 8.59E−05 | + |
| TIDAL-total | Normalized coverage | 5.03E−03 | + | 2.17E−07 | + |
| TIDAL-weighted | Normalized coverage | 2.12E−01 | + | 4.75E−05 | + |
| TEMP-total | Normalized coverage | 2.83E−01 | + | 4.97E−01 | + |
| TEMP-weighted | Normalized coverage | 9.28E−01 | − | 3.66E−01 | − |
| HISAT/Cufflinks CN | Normalized coverage | 5.63E−01 | − | 2.15E−04 | + |
| HISAT/StringTie CN | Normalized coverage | 3.74E−01 | − | 9.43E−04 | + |
| gsnap/Cufflinks CN | Normalized coverage | 3.32E−01 | − | 3.44E−04 | + |
| gsnap/StringTie CN | Normalized coverage | 1.34E−02 | − | 5.96E−02 | + |
| STAR/Cufflinks CN | Normalized coverage | 7.37E−05 | − | 1.19E−02 | − |
| STAR/StringTie CN | Normalized coverage | 3.61E−04 | − | 8.72E−01 | + |
| HISAT/Cufflinks FL-CN | Normalized coverage | 3.99E−05 | − | 6.77E−01 | − |
| HISAT/StringTie FL-CN | Normalized coverage | 5.10E−07 | − | 8.41E−04 | − |
| gsnap/Cufflinks FL-CN | Normalized coverage | 1.03E−01 | − | 1.47E−01 | + |
| gsnap/StringTie FL-CN | Normalized coverage | 3.21E−01 | − | 1.68E−01 | + |
| STAR/Cufflinks FL-CN | Normalized coverage | 1.32E−03 | − | 7.46E−01 | − |
| STAR/StringTie FL-CN | Normalized coverage | 3.65E−04 | − | 8.54E−01 | + |
| HISAT/Cufflinks FP/KP | Normalized coverage | 2.23E−01 | − | 7.24E−01 | − |
| HISAT/StringTie FP/KP | Normalized coverage | 1.02E−02 | − | 4.57E−01 | − |
| gsnap/Cufflinks FP/KP | Normalized coverage | 7.46E−02 | − | 4.47E−01 | − |
| gsnap/StringTie FP/KP | Normalized coverage | 9.79E−02 | − | 6.16E−01 | − |
| STAR/Cufflinks FP/KP | Normalized coverage | 1.06E−06 | + | 7.75E−08 | + |
| STAR/StringTie FP/KP | Normalized coverage | 7.08E−01 | − | 2.54E−01 | + |
The fit of null models that include hobo abundance as a fixed effect and experimental block as a random effect was compared to alternative models that also include an estimate of P-element abundance or structural variation. The P-values reported result from a likelihood ratio test comparing the two model fits, with a value of <0.05 indicating that the model including an estimate of P-element abundance or structural variation provided a significantly better fit to the data than the null model. For each model, the direction of the association between paternal P-elements and hybrid dysgenesis is also indicated (±). All model pairs were fitted to both the full data set and the data set excluding the outlier strain RAL-42. Complete information about model components, fit, and comparison can be found in Table S8. CN, copy number; FL, full-length; FP, full-length P-element.