Announcement

Collapse
No announcement yet.

Counting reads in miRNA experiment

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Counting reads in miRNA experiment

    Counting reads in miRNA experiment

    Dear all,
    in our lab we are running an Illumina small RNA protocol on human sample to collect miRNA sequence and I am running the bioinformatics part.
    During library preparation we are using kit that are unstreanded. After sequencing (Illumina single read) for each sample I got the majority of reads with 5-3' orientation (about 90% of the reads) and "few" reads with orientation 3'-5'.
    I have cleaned my raw data removing low quality reads, illumina adapter, reads longer than 27 nt and reads shorter than 17 nt after adapter removing. I aligned to human genome and now I need to assign and count my reads to microRNA. For doing that I am using FeatureCounts (suite subread) and I am using the gff file downloaded from mirBase. I have a doubt in summarizing the reads: do I need to count the reads independently from their orientation? Do I need to sum the reads with orientation 5'-3' and 3'-5'? Or do I only need to count the reads with 5'-3' orientation?
    Many thanks,
    Sara

  • #2
    Hi,
    If the protocol you follow keeps no strand information of RNA ("unstreanded") then I think you should consider no orientation for counting; sum all 5'-3' and 3'-5' reads for the same features. I haven't work with FeatureCounts, with htseq-count (http://www-huber.embl.de/users/ander...doc/count.html) it would be with the --stranded=no option.
    Last edited by cascoamarillo; 01-15-2016, 12:37 PM.

    Comment


    • #3
      Hi,
      many thanks for your suggestion.
      Best,
      Sara

      Comment


      • #4
        With featureCounts, the default setting is unstranded read counting (-s 0).

        If you performed unstranded sequencing, the percentage of sense strand reads should be similar to that of antisense strand reads. You may need to check if your sequencing was actually stranded.

        Comment

        Working...
        X