HS_LIBRARY_SIZE vs. ESTIMATED_LIBRARY_SIZE

[sgroenewold]

Nothing came up on a search here today for me, so I dug into these Picard modules.

Looks like MarkDuplicates.ESTIMATED_LIBRARY_SIZE estimates the number of unique fragments supplied by the entire library, while HsMetrics.HS_LIBRARY_SIZE focuses only on the unique/duplicate reads aligned to the bait regions. Both use the Lander-Waterman equation to work backwards from unique reads vs. total reads.

Does that assessment agree with others here?

[Steve_DePalma]

That has always been my understanding, yes.