pop from empty list error from feature-classifier

[Xio_Lee]

Hi,

I encountered some error after training silva 132 release for 16s and testing the classifier. Checked with this post: [qiime2 pop from empty list]. However, I checked with a python script, did not get a line with missing taxonomy.

def checkTaxa(file):
    checkList = ["D_0__", "D_1__", "D_2__", "D_3__", "D_4__", "D_5__", "D_6__", ]
    with open(file, 'r') as f:
        for line in f:
            line = line.strip()
            if not any(n in line for n in checkList):
                print(line)

Not sure what the issue was.

I am using qiime2 2019 04

for input sequence, I used the 99 fasta and with the 7 levels taxonomy. I stripped off "#" in the taxonomy as well as the white space.

attached are the debug log and scripts.
qiime2-debug.txt (3.2 KB)

train_classifier.txt (1.4 KB)
test_classifier.txt (492 Bytes)

Thank you

[Nicholas_Bokulich]

Hi @Xio_Lee,

@Xio_Lee

for input sequence, I used the 99 fasta and with the 7 levels taxonomy. I stripped off “#” in the taxonomy as well as the white space.

The modifications you made to the taxonomy file must have somehow resulted in uneven levels.

Note that your checkTaxa function will not actually validate this appropriately:

@Xio_Lee

if not any(n in line for n in checkList):

You should check for all instead of any (though theoretically that may not catch missing levels if, e.g., you are missing a semicolon but have the rank label).

Give that a try and let us know if that helps track down the issue!