mouse mm9 miRNA in gtf format

[roll]

Hi,
Where can i find mouse miRNA mm9 reference data in gtf format?
I know mirbase but i could only found these in gff format.
I would like to use this for htseq-count.

[A.N.Other]

Ensembl will get you download whole legacy GTFs from their FTP (ftp://ftp.ensembl.org/pub/ --> choose the correct release --> gtf --> Mus musculus). You can filter out the miRNAs from the whole list, but if you're using htseq-count it might just be easier to give it the complete GTF and do the filtering afterwards.

[roll]

Ensembl will get you download whole legacy GTFs from their FTP (ftp://ftp.ensembl.org/pub/ --> choose the correct release --> gtf --> Mus musculus). You can filter out the miRNAs from the whole list, but if you're using htseq-count it might just be easier to give it the complete GTF and do the filtering afterwards.

What I am mising to understand is that my boss wants to have a table like the following - for each mirna
Cond1 Cond2 Cond3
mirna1 3 5 7
mirna2 2 0 6
…

When I download the gene.gtf from ensemble I only got these counts gene wise.

How can i obtain the counts for each mirna on 3 conditions (like the table i mentioned above).

[GenoMax]

You are going to use "HTSeq-count" with the GFF file to get that table: http://www-huber.embl.de/users/ander...unt.html#count after you filter the GFF file leaving only miRNA's like A.N.Other suggested.

This is how you can do that from a unix machine. Get the right GTF file for the build you used for your alignments.

$ wget ftp://ftp.ensembl.org/pub/release-74/gtf/mus_musculus/Mus_musculus.GRCm38.74.gtf.gz

$ gunzip Mus_musculus.GRCm38.74.gtf.gz

$ cat Mus_musculus.GRCm38.74.gtf | grep "miRNA" > file_name_with_miRNA.gtf

$ cat Mus_musculus.GRCm38.74.gtf | grep "snoRNA" > file_name_with_snoRNA.gtf

$ cat Mus_musculus.GRCm38.74.gtf | grep "lincRNA" > file_name_with_lincRNA.gtf

$ cat Mus_musculus.GRCm38.74.gtf | grep "snRNA" > file_name_with_snRNA.gtf

You can then use http://www.sequenceontology.org/cgi-bin/converter.cgi to convert the gtf files to gff.

[roll]

You are going to use "HTSeq-count" with the GFF file to get that table: http://www-huber.embl.de/users/ander...unt.html#count after you filter the GFF file leaving only miRNA's like A.N.Other suggested.

This is how you can do that from a unix machine. Get the right GTF file for the build you used for your alignments.

$ wget ftp://ftp.ensembl.org/pub/release-74/gtf/mus_musculus/Mus_musculus.GRCm38.74.gtf.gz

$ gunzip Mus_musculus.GRCm38.74.gtf.gz

$ cat Mus_musculus.GRCm38.74.gtf | grep "miRNA" > file_name_with_miRNA.gtf

You can then use http://www.sequenceontology.org/cgi-bin/converter.cgi to convert the gtf file to gff.

My understanding is that if i do the above, i will get the miRNA count per gene. Whereas i am interested in is counting each miRNA type in each sample, like the table above. Is it possible to do that with htseq-count?

[dpryan]

How easy this is depends on how the annotation is constructed. htseq-count (or featureCounts, for a generally quicker alternative) will count according to whatever feature you tell it to. So, if your annotation has a field that specifies the miRNA type, then htseq-count can be told to count according to that.

One caveat is that it will still ignore multimappers, which may be quite prevalent in your situation. The proper handling would be to increment the count of the feature by one if all of the multiple mappings of a particular read fall only in a single feature. I don't think htseq-count will do that for you.

[dpryan]

I should add that featureCounts might allow this (at least Wei Shi has mentioned in the past that this is the case).

[GenoMax]

HTSeq will not make that table for you. After taking into consideration Devon's suggestion you can easily create the table yourself once you have the counts for each sample.

[sudders]

I believe the the Kraken pipeline will process miRNA-seq data into the format you require - it includes options to summarise counts per gene or per mature miRNA.

[roll]

HTSeq (or featureCounts) will not make that table for you. After taking into consideration Devon's suggestion you can easily create the table yourself once you have the counts for each sample.

i used featureCounts and used gff file from mirBase. After selecting the right options, it returned me format i wanted it to be (count of each miRNA in each sample). Do you think, there is a mistake?

[GenoMax]

I should not have included featureCount in post #8. My apologies (post amended).

Devon (post #7) had said that featureCounts may do this (and it indeed seems to). He has more experience with analysis.