Closed mortonjt closed 7 years ago
ping on this issue - this is a serious user issue, because it is a major technical barrier when trying to perform taxonomic classification. Otherwise the user will be forced to be running
tr '[:lower:]' '[:upper:]'
At the very least, we need to document this.
I too have encountered this problem, however I assumed this was a sequencing artifact (low quality NTs). There's an issue for this in q2-types https://github.com/qiime2/q2-types/issues/91 That being said, if deblur is changing NTs to lowercase, what is the rationale for that?
Yoshiki Vázquez-Baeza
On May 30, 2017, at 6:22 PM, Jamie Morton notifications@github.com wrote:
ping on this issue - this is a serious user issue, because it is a major technical barrier when trying to perform taxonomic classification. Otherwise the user will be forced to be running tr '[:lower:]' '[:upper:]'
At the very least, we need to document this.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
It's stemming from MAFFT iirc. I'd be :+1: for a PR to coerce to all upper case
On Wed, May 31, 2017 at 6:40 AM, Yoshiki Vázquez Baeza < notifications@github.com> wrote:
I too have encountered this problem, however I assumed this was a sequencing artifact (low quality NTs). There's an issue for this in q2-types https://github.com/qiime2/q2-types/issues/91 That being said, if deblur is changing NTs to lowercase, what is the rationale for that?
Yoshiki Vázquez-Baeza
On May 30, 2017, at 6:22 PM, Jamie Morton notifications@github.com wrote:
ping on this issue - this is a serious user issue, because it is a major technical barrier when trying to perform taxonomic classification. Otherwise the user will be forced to be running tr '[:lower:]' '[:upper:]'
At the very least, we need to document this.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/biocore/deblur/issues/148#issuecomment-305189582, or mute the thread https://github.com/notifications/unsubscribe-auth/AAc8svXiE0dhnFpptMGmPg00BC60sK-lks5r_W21gaJpZM4NCSce .
Just stumbled across this through the q2-types
issue. One thing QIIME 2 does (in q2_alignment.methods.mafft
) is pass --preservecase
to MAFFT, which may be an option for this as well by adding it to here.
Which q2types issue?
On Jun 15, 2017 12:48, "Jorden Kreps" notifications@github.com wrote:
Just stumbled across this through the q2-types issue. One thing QIIME 2 does (in q2_alignment.methods.mafft) is pass --preservecase to MAFFT, which may be an option for this as well by adding it to here https://github.com/biocore/deblur/blob/afba6a195bf6eb2c249bccffc3a842da64fdfaca/deblur/workflow.py#L519-L520 .
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/biocore/deblur/issues/148#issuecomment-308847329, or mute the thread https://github.com/notifications/unsubscribe-auth/AAc8smH03OCHImeaOYaSKpmbkKaGUSxgks5sEYqDgaJpZM4NCSce .
The importing issue that Yoshiki mentioned, about QIIME 2 being unable to import lowercase sequences. I just ran into an issue with that by running MAFFT external to q2, and since he tagged it here this issue showed up on that page.
Ah, kk
On Jun 15, 2017 12:52, "Jorden Kreps" notifications@github.com wrote:
The importing issue that Yoshiki mentioned, about QIIME 2 being unable to import lowercase sequences. I just ran into an issue with that by running MAFFT external to q2, and since he tagged it here this issue showed up on that page.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/biocore/deblur/issues/148#issuecomment-308848238, or mute the thread https://github.com/notifications/unsubscribe-auth/AAc8sliAMGnBID96x44WW2VjMKN4Tpudks5sEYuAgaJpZM4NCSce .
Any way this can be included in a release soon?
Here is an example of deblurred sequence
Is there a reason why this gets outputted? This can cause problems downstream, since q2-feature-classifier cannot handle lowercase characters.