Classifier just class level

[110147]

I recently do classifier meet some problem , when I do taxonomy for my smple I usually get the result juse a class level ,like the picture ,I find the reads usually very long ,so I usually set the -trim-p is 0 and the -trun-p is 400 ,then I will get the long DNA, but I can't get the result about other levels ,I meet the suitation three times ,So I hope you can help me and tell me the reson , I need to solve the problem ,thank you very much .

[Nicholas_Bokulich]

Hi @110147,
Looks like you are only receiving kingdom-level classifications. This is almost always an indication of human error, specifically whether the correct reference database or classifier was used. Please answer the following questions:

1. What gene target/primers are you using?
2. What classification method are you using? Please post the exact commands.
3. I see that you are using greengenes. Did you use one of QIIME2’s pretrained classifiers or are you training your own classifier?

Chances are either:

1. you are using a classifier trained with the wrong primers
2. greengenes is not appropriate for your data. Are these bacterial 16S rRNA gene reads?

If you are receiving non-kingdom classifications, see this post for further troubleshooting.

I hope that helps!

[110147]

thank you reply my question , I also try many times and the last I get the same result which just have the kingdom level ,the data from the NCBI and I use the article data ,I use the V1-2 primers and I use the greenenes classifically ,and I use the qiime2 offer database ,I want to know the train database is depend on my primers .and then I want to know maybe I should build the different primers and usually u se database for my analyse. and I use the silva and the greenene database do the analyse ,but I don’t know this is ture .so I should train the silva and greengene database ,and I should train in different orimers ? thank you very much.

[110147]

thank you ,I want to know if I have other target/primers , I want to train my database and I want to know which "ref-seqs.qza " I can use ,in the wed https://docs.qiime2.org/2018.2/data-resources/ just have the full and the 515F/806R primers , I want to know I should use which to train my database ,and if I use the silva database ,I use the 99.otus.fastas and the majority_taxonomy_7_levels.txt or consensus_taxonomy_7_levels.txt ,I don’t know which is true ,but the last I find the result isn’t good ,,yes, the result just little is annotated ,and I use the greengene have many levels .so , I want to consult the team about the classifier which is the best annotation way . thank you .

[thermokarst]

@Nicholas_Bokulich

If you are receiving non-kingdom classifications, see this post for further troubleshooting.

Hi @110147, did you see this link that @Nicholas_Bokulich recommended you visit? I will print it here again for you:

Train you classifier - taxonomy

Hi @Jasmine84, See this post. Depending on the length of your sequences, the primers you are using, etc, deeper than order level might be unreliable. On 250 bp 16S rRNA gene sequences, we can usually get genus level and some species with a good degree of reliability. The naive bayes classifier in QIIME2 is designed to be cautious and reliable, to avoid false-positive errors, but you can lower the confidence parameter to make it less cautious. If you have: only kingdom-level classification, …

As @Nicholas_Bokulich asked above, can you please post the exact commands you ran?