Closed ArtPoon closed 8 years ago
For each `good' patient in the lanl dataset I calculated the percentage of DNA samples that were below the GLM calculated from the plasma samples.
Good means that there were no processing errors and that the GLM slope converges with positive.
When I used:
This would suggest - though I feel not strongly - that the DNA samples are less evolved.
Binomial regression (unsigned test) where probability parameter expresses whether HIV DNA is more or less evolved than expected.
This does not appear in the manuscript.