However, since previous immunoproteogenomics studies were based on low-coverage 454 technology, it was not clear whether this lack of correlation represents sampling artifacts or reflects the fact that previous immunoproteogenomics studies did not correct sequencing errors.
While our understanding of some of the genetic underpinnings of DIPG has improved over the last few years, the assumption has been that these represent brainstem GBMs and that low-grade histology seen in biopsy specimens represents sampling bias.
Most likely this finding represents sampling bias; microorganisms sent from the districts are less likely to survive than samples taken in Nuuk, and district doctors may be less likely to submit samples than Nuuk doctors, given the longer time to receive test results.
To account for the inflationary effect of sampling variance, we considered the model logEC10 = μ + ε, where μ is the underlying true (unknown) logEC10 and ε represents sampling variation.
