# GFA2bin tool

1. Below is "our" pipeline.

2. The tool was developed in the meantime: https://github.com/MoinSebi/gfa2bin
This tool didn't start from our input and output. It worked directly on the matrix of reads per node. As we have done in the past, it created a matrix of genotypes based on the matrix of reads per node (using different normalization steps).


Below is how it works:

`gfa2bin cov --p file.pack --output <output>`


Normalizations:

 -a, --absolute-threshold <absolute-threshold>    Set a absolute threshold

--method <method>        Normalization method (mean|median|percentile)

A hard cutoff value for coverage--> absolute threshold

If I tried it with the example tests, it didn't work.

Main differences:

- Compared to our pipeline, it used `gaf2pack` to produce the matrix of reads per node, which gives this output:
With our tool, we obtained instead node.ids and coverage only.

- It used a rGFA file, https://github.com/lh3/gfatools/blob/master/doc/rGFA.md

- In addition, it's possible to compress the pack file (maybe interesting for human data):
Compress a plain-text coverage file to "pack compressed". Mainly used to reduce the storage size of the coverage file. Maximum coverage in these files is 6553. Higher coverages are truncated.

#### One sample

- Use `gaf2pack`
`gaf2pack --gfa D_C_mm10.fa.gz.bf3285f.eb0f3d3.867196c.smooth.final.gfa --alignment BXD32_BDref_inject.gaf --output BXD32.gaf2pack.txt`

- Use `gfa2bin`

`gfa2bin cov --output /scratch/BXD32.geno --packlist /lizardfs/flaviav/mouse/panQTL_result/BXD32/path.txt`

Default:

4740 entries have been truncated (have a coverage above 65,535).
4740 entries have been truncated (have a coverage above 65,535).