If several studies on a marker are categorized as initial studies, we use the sign of the most extreme result.
Within Table 2, each of the 3 analysed markers was categorized as a dichotomous variable and a correlation between selected baseline patient characteristics and molecular marker results was performed.
Tissue expression of markers was categorized dichotomously as either 'over-expression' or 'under-expression', according to the 'quick score' method [ 17, 18], which multiplies the percentage of stained cells (P) by the intensity of staining (I).
A total of 77 (or 80.2%) markers were categorized as good markers (Table S7).
For continuous variables, markers were categorized into a low and high group using the median value as the break point.
Candidate SNP markers were categorized as testcross in pair-wise comparisons of genotypes, whether a heterozygous imputation was present for one parent only (testcross) and a homozygous site was predicted for the other.
When the markers were categorized by the tertile distributions of the control group, the highest tertile of FT4 (OR = 1.73, 95% CI = 1.11 - 2.69) showed an increased risk of thyroid cancer, while the middle tertile of TSH (OR = 1.77, 95% CI = 1.14 - 2.74) was associated with thyroid cancer risk.
When the markers were categorized by the tertile distributions of the control group, the highest tertile of FT4 (OR = 1.73, 95% CI = 1.11 - 2.69) and the middle tertile of TSH (OR = 1.77, 95% CI = 1.14 - 2.74) were associated with an increased risk of thyroid cancer by multivariate analyses.
Risk marker information was categorized into common domains, and risk markers were consolidated where possible (not shown).
The relative intensity of PL2/3 (Fig. 5F) and #32 (Fig. 5G) nucleosome markers in the doughnut was categorized in two groups: more intense or equally intense nucleosome signal in the doughnut and less intense or no H3.1/3.2 signal in the doughnut.
