ValueError: invalid DNA character in sequence: N

[dchandran]

Many genomic fasta files in genbank have N in the sequences, and for these sequences, I am getting the "ValueError: invalid DNA character in sequence: N" error when using "sourmash compute". We would like to build signature files automatically from genbank genomes; since sourmash does not allow N in the sequence, this is somewhat of a blocking issue.

[luizirber]

Hello @dchandran,

what sourmash version are you using? It seems you're using sourmash 1, did you install it with pip? If this is the case, can you run
pip install --pre sourmash instead (and then you get the latest version, 2.0.0a10).

This example uses a small viral genome that has N bases (and using sourmash==2.0.0a4). Can you reproduce the error with this command?

curl https://ftp.ncbi.nlm.nih.gov/genomes/all/GCA/001/857/745/GCA_001857745.1_ASM185774v1/GCA_001857745.1_ASM185774v1_genomic.fna.gz | zcat - | sourmash compute - -o test.sig

[ctb]

(Linking in the relevant issues so we can trace this in the future better :)

See also #40, #192, and especially #148. We do the same thing that mash does! marbl/Mash#46!

[ctb]

Closing for now.