Mapping to miRBase

Caroline Plus · 12-21-2017, 12:08 PM

Hi guys,

How to do the mapping to miRBase to identify human miRNAs?

1. At first you should download 2 fasta files from the website: http://www.mirbase.org/ftp.shtml (and merge them into one file):
hairpin.fa Fasta format sequences of all miRNA hairpins
mature.fa Fasta format sequences of all mature miRNA sequences

3. Then, you should generate gtf for Bowtie 1 (if you want to use specifically Bowtie 1) or gff for Htseq (based on the downloaded fasta files).

I did the analysis this way, but I received very strange results (for example mir-8689, mir-7689 - I mean miRNAs with very high numbers) instead of what I was expecting (mir9, mir31, etc.- miRNAs with much smaller numbers).

Do you know maybe what I could do wrongly?

Thanks for your answer in advance!

sbarberan · 01-05-2018, 04:43 PM

Hi Caroline,

What I do is download the sequences from miRBase and then change all 'U' to 'T' and then create the index with bowtie-build and then align.

Could it be that because you are not changing U to T you only get hits for miRNAs with no Us?

Cheers,
sergio

Caroline Plus · 01-07-2018, 08:38 PM

Hello Sergio,

Thanks so much! I have just tried your approach! You are right! It is working!

Best wishes,
Caroline

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
almost no hits when mapping against miRBase	Irina Pulyakhina	Bioinformatics	0	09-04-2015 07:15 AM
miRBase - Why do they not keep name conventions?	Soeren_Mueller	Bioinformatics	0	08-20-2014 08:34 AM
In miRBase search	unique379	Bioinformatics	1	07-09-2013 01:59 PM
Mapping miRNA on miRBase	nedoluzhko	Bioinformatics	11	09-16-2012 01:25 AM
miRNA-seq - mapping to MIRBASE	hrajasim	Illumina/Solexa	0	02-28-2010 04:29 PM

12-21-2017, 12:08 PM	#1
Caroline Plus Junior Member Location: Spain Join Date: Oct 2015 Posts: 9	Mapping to miRBase Hi guys, How to do the mapping to miRBase to identify human miRNAs? 1. At first you should download 2 fasta files from the website: http://www.mirbase.org/ftp.shtml (and merge them into one file): hairpin.fa Fasta format sequences of all miRNA hairpins mature.fa Fasta format sequences of all mature miRNA sequences 3. Then, you should generate gtf for Bowtie 1 (if you want to use specifically Bowtie 1) or gff for Htseq (based on the downloaded fasta files). I did the analysis this way, but I received very strange results (for example mir-8689, mir-7689 - I mean miRNAs with very high numbers) instead of what I was expecting (mir9, mir31, etc.- miRNAs with much smaller numbers). Do you know maybe what I could do wrongly? Thanks for your answer in advance!

01-05-2018, 04:43 PM	#2
sbarberan Member Location: Santa Cruz, CA Join Date: Feb 2017 Posts: 17	Hi Caroline, What I do is download the sequences from miRBase and then change all 'U' to 'T' and then create the index with bowtie-build and then align. Could it be that because you are not changing U to T you only get hits for miRNAs with no Us? Cheers, sergio

01-07-2018, 08:38 PM	#3
Caroline Plus Junior Member Location: Spain Join Date: Oct 2015 Posts: 9	Hello Sergio, Thanks so much! I have just tried your approach! You are right! It is working! Best wishes, Caroline