inner -> outer join with imputation

[rhagenson]

Right now the paired counts are inner joined (the default behavior of merge).

crispr-screening/bin/extract-reads.py

Lines 75 to 85 in 1a6f8c1

	finalDf = initialDf.merge(
	finalDf, on=["id", "gene"], suffixes=["_initial", "_final"]
	).rename(
	columns={
	"id": "sgRNA",
	"gene": "Gene",
	"count_initial": sample + ".initial",
	"count_final": sample + ".final",
	}
	)
	finalDf.to_csv("count-" + sample + "-i.f.csv", index=False)

However, the downstream analysis (i.e., MAGeCK mle) should be able to handle merging counts properly if we do a full (outer) join then impute with ones/1s.

[rhagenson]

This is not exactly necessary right now as usually all samples are run at once so any sgRNAs found in a single sample will get a count of 1 for samples where it was not found. However, this implicit application will not hold if #4 is implemented in certain ways. It best to make the switch described here as a bit of defensive programming (#2).