vcf_compare.md

filter_vcf cann be chained into diff (well, via piping) | any results can be seen as counts, which is len(df)

Did we get the same var calls? Does our method introduce more/less ambiguity

vcf_compare ambiguous_bases . filter_vcf depth --threshold FILE . filter_vcf [CBD --minimum MIN]... [--count | -c] FILE (#allow no filter) . filter_vcf [ALT --exists] . diff ALT [--threshold THRESH] FILE1 FLIE2 . mutations (# this creates a sub-df which can then be dfd.)

Script will expect The following arguments:

1. two .vcf files
2. those fields to check for differences on (i.e. Called base, etc.). Optional, check all fields by default
3. A numerical (percent-based?) thershold for deciding if two attributes are different. Optional, 0.01 by default i.e. CBD at position 0 is: file1: 12 file2: 14 if threshold is 2 or less, than the record at position 0 will be included under the CBD heading in the diffs.txt file.

I propose the ouoptut will be in the form of the following files: sumarry.txt

Statistical summary of the differences between the two files, like average and total call depth difference, etc.

diffs.txt A field by field listing of how many different records appeared for each field, followed by the differing vcf records (as seen in the original files) listed with their position. i.e., file1 has CBD=99 at position 0, while file2 has CBD=150 at position 0, similar at position 9

There were 2 records with differing called base depth with threshold 10:
0> VCF record from file 1
0< VCF record from file 2
9< VCF record from file 1
9> VCF record from file 2

headers.txt The ouptut of diffs.txt without the actual records listed. i.e.:

There were 2 records with differing called base depth with threshold 10:
There were 4 records with differing called bases.

diffbases.diff The output of the existing vcf_diff.py script (showing where Called bases were different from references) piped to diff. Or, these records listed side-by-side, as part of diffs.txt.