Phanta now has many databases and it is getting clunkier to try to bundle it into a single stage. Lets split this back out.
an alternative would be to add the genome size files to all the kraken databases we use, run kraken/bracken against a list of databases, and run the genome size correction script from phanta on all
Phanta now has many databases and it is getting clunkier to try to bundle it into a single stage. Lets split this back out.
an alternative would be to add the genome size files to all the kraken databases we use, run kraken/bracken against a list of databases, and run the genome size correction script from phanta on all