Announcement

Collapse
No announcement yet.

Normalize small RNA data with spike-in controls?

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

  • Normalize small RNA data with spike-in controls?

    Dear all,

    I am planning to sequence small RNAs from my control and mutant cells and I was wondering what is the best way to normalize the data. My initial data shows that the small RNAs in wild-type are coming from the entire genome (sequencing) and the mutants show reduced levels of small RNAs (based on PAGE gel analysis from equal amounts of RNA). However, if the mutants make all small RNAs proportionally less, how do I normalize? Has anyone had luck using small control RNAs as spike-ins? My concern is that the sequence of the control RNA could control ligation efficiency and have some potential sequencing bias.
    Additionally, can someone comment on whether I absolutely need biological duplicates for comparisons? Can DEseq or EdgeR packages be used without replicates?

    Thanks in advance

  • #2
    Originally posted by jazz View Post
    However, if the mutants make all small RNAs proportionally less, how do I normalize? Has anyone had luck using small control RNAs as spike-ins?
    That's a difficult problem. I'm suspect there might not be a good solution, but you might look at the RUVseq package and accompanying paper.

    Originally posted by jazz View Post
    Additionally, can someone comment on whether I absolutely need biological duplicates for comparisons? Can DEseq or EdgeR packages be used without replicates?
    In my opinion, the expense of doing biological replicates will repay you many times over. However look at the section "What to do if you don't have replicates" in the edgeR User's Guide.

    Comment


    • #3
      Thanks. I will check the RUVseq. Can I use RPKM for my case?

      Comment


      • #4
        Originally posted by jazz View Post
        Can I use RPKM for my case?
        You cannot use RPKM with either edgeR or DESeq, as the documentation for those packages will tell you.

        All the packages (edgeR, DESeq, RUVseq) use read counts, which is what you should get when you process your data anyway.

        Comment


        • #5
          We are also planning some sRNA-seq in which we expect large differences in piRNA expression, and are thinking about these spike-ins. However we are still undecided and I would like to hear what experiences people have with this.

          Comment

          Working...
          X