ecogenomics / mingle Goto Github PK

It would be helpful if the pre- and post-genes reported by mingle indicated their orientation on the contig. This would allow assessment of conserved gene order and orientation across genomes.

@dparks1134 I'm guessing 500 is the number of blast hits rather than the number of hits in the database. Maybe just use a bigger number of hits returned by default?

uqbwoodc@mrca003:20150220:/srv/projects/abisko/shotgun_abundance/53_doe_talk_functional_gene_pca/bclA$ mingle blast -q AAN32623.1.fasta -o AAN32623.1.mingle --cpus 40
[2015-02-20 22:49:49] INFO: Reading metadata information from ARB GreenGenes file.
[2015-02-20 22:49:49] INFO: Reading taxonomy information from ARB GreenGenes file.
[2015-02-20 22:49:49] INFO:   Found records for 15134 genomes.
[2015-02-20 22:49:49] INFO: Identifying homologous genes using BLAST.
[2015-02-20 22:50:17] INFO:   Identified 500 homologous genes.
[2015-02-20 22:50:17] INFO: Extracting homologous sequences.
[2015-02-20 22:51:10] INFO: Creating GreenGenes-style file for ARB.
[2015-02-20 22:51:11] INFO: Inferring multiple sequence alignment.
[2015-02-20 22:53:20] INFO: Inferring gene tree.

There are a number of characters that are invalid within user supplied sequences. It would be good to do an initial check for these characters: ,;()|~.

Unfortunately, genomes have not been added to the genome tree database in a consistent matter. As such, the metadata is not entirely consistent and identification of public genomes using entry['core_list_status'] == 'public' may result in some public genomes being missed. If we appear to be missing some expected genomes when filtering on the public flag this may be why.

Should we just filter on 'A' vs. 'C' genomes? Or just let this go for now with the idea that the GTDB will improve with time?

commit
geronimp@31e5d2c