Skip to main content
. 2021 Apr 19;2(2):100483. doi: 10.1016/j.xpro.2021.100483

Table 1.

Regression models for detecting ancestry-associated molecular correlates

Alternative hypothesis Method Example model
Somatic alteration is associated with local ancestry Logistic Regression somatic alteration ~ local ancestry + percentage of EUR ancestry + percentage of AFR ancestry
Ancestry is associated with arm-level SCNA Logistic Regression ancestry ~ arm-level SCNA + aneuploidy + genome doubling + age + gender + subtype
Ancestry is associated with TMB, aneuploidy, and IMS Logistic Regression ancestry ~ aneuploidy + TMB + IMS + age + gender + subtype
DNA methylation is associated with ancestry Linear Regression DNA methylation level (β value) ~ ancestry + tumor purity + age + gender + subtype
mRNA expression is associated with ancestry Linear Regression log2(RPM) ~ ancestry + TCGA plate number + subtype
miRNA expression is associated with ancestry Linear Regression log2(RPM+1) ~ ancestry + subtype
Pathway activity is associated with ancestry Linear Regression IPL ~ ancestry + subtype

RPM: Read Per Million. IPL: Integrated Pathway Level.