Pathway coverage vs. pathway completeness

Hello! Sorry if this question about concepts is a little basic, I am still a student.

One of Humann2 outputs is pathway coverage, which is dependent on reaction detection confidence (“a pathway with coverage = 1 is considered to be confidently detected, all of its member reactions were also confidently detected”). I also know, from other works, that there is a measure called completeness, which is the fraction of its genes that are present (100% completeness implies that all genes of the pathway are present).

Given that the presence of a reaction relies on the presence of genes, and that the presence of a gene requires its confident detection, which is the practical difference between pathway completeness and Humann2’s pathway coverage?

Thank you in advance.

Sorry for the late reply - coverage requires there to be a complete path to the pathway’s products using only confidently detected reactions, but allowing up to 1 reaction to be missing (gap-filling) and optimizing over subpathways (so, for example, if a pathway could start from reaction A OR B, then HUMAnN would only be looking for at least one of these reactions to be present). “Completeness” (per your definition) sounds like a more straightforward accounting of which fraction of total reactions annotated to the pathway were detected (non-zero).

2 Likes