in spacegraphcats, we are consuming the output of your excellent tool and doing (or attempting to do) Clever Graph Things with it. (thank you for bcalm!) however, we are running into a problem that in some cases there are nodes u, v where node u has an edge to v but node v does not have an edge to u.

more specifically, we see a number of examples like the following two headers in the unitigs output:

>206849 LN:i:41 KC:i:96 km:f:8.7  L:-:55026:- L:-:69342:-  L:+:103808:- L:+:203510:-
...
>69342 LN:i:134 KC:i:2260 km:f:21.7  L:+:69614:- L:+:155497:-  L:-:61581:+ L:-:167785:+
...

which show that 69342 is in the edge list of 206849, but 206849 is not in the edge list of 69342.

We have code to detect this here,

spacegraphcats/spacegraphcats@dc9a9cc

and I am happy to provide full execution instructions should you wish (should be: checkout that branch; pip install -r requirements.txt; make twofoo.fq.gz; conf/run twofoo build)

your thoughts appreciated!

output from latest bcalm master is here: https://osf.io/zgx3r/ (generated with bcalm built from commit c41f70f on my mac os x laptop)

input files are generated from the rules here,

briefly

curl -o akker-reads.abundtrim.gz -L https://osf.io/dk7nb/download
curl -L 'https://osf.io/7az9p/?action=download' > shew-reads.abundtrim.gz
gunzip -c shew-reads.abundtrim.gz akker-reads.abundtrim.gz | gzip -9c > twofoo.fq.gz

and then the command line execution is:

bcalm -in bcalm.twofoo.k31.inputlist.txt -out twofoo/bcalm.twofoo.k31 -kmer-size 31 -abundance-min 1