When analyzing formalin-fixed paraffin embedded tissues, one can expect considerable rates of low-frequency C>T and G>A artifacts (see Do and Dobrovic, Clinical Chemistry 2014).
The variant calling grammar allows to model them as a separate event, allowing to filter them away while considering the involved uncertainty.
As can be seen, here we introduce an additional type of atomic expression, selecting particular substitutions (e.g. C>T
, allowed is the IUPAC DNA alphabet).
samples:
ffpetumor:
universe: "[0.0,1.0]"
resolution: 0.01
expressions:
ffpe_subst: "C>T | G>A"
events:
ffpe_artifact: "($ffpe_subst) & ffpetumor:]0.0,0.05["
present: "(($ffpe_subst) & ffpetumor:]0.05,1.0]) | (!($ffpe_subst) & ffpetumor:]0.0,1.0[)"