miRNA-seq DE tool for using with normalized (by mapping) count data?

[PabloMarin-Garcia]

From the docs of DESeq and forums seems to me that it is not a good idea to use data count that has been normalized by multiple mapping. My read counts has been produced with illumina's flicker that produce HNA (Hit Normalize Abundance), therfore not the right input for DESeq. Then I have to options:

a) use another tool for the DE that can use this type of counts. Any suggestion?

b) process again the sequended reads to obtain raw data. Do someone know how to tell Casava to do the mapping in a way that flicker gives raw data?. Would you suggest another pipeline for processing the miRNA that gives the right input for DESeq?


-Pablo

[sjm]

E-miR program

Try taking a look at the article below. I've been using this pipeline to analyze miR sequencing reads generated via Illumina's TruSeq small RNA preparation kit, on an Illumina HiSeq. The software supplied by the authors uses Illumina's Eland aligner (which I don't have) or bowtie (which is freely available). Note, however, that the perl script supplied in the 'bowtie' version of their package doesn't allow for multiple alignments (i.e. it throws out reads aligning to multiple places in the genome, which throws out miRs from more than 1 genomic locus, which is silly). It's simple enough to modify their bowtie command.

The package gives raw miR counts per sample as well as normalized reads per million mapped - but the raw count data can easily go into DESeq or edgeR.

BMC Genomics. 2010 Dec 20;11:716.
New methods for next generation sequencing based microRNA expression profiling.
Buermans HP, Ariyurek Y, van Ommen G, den Dunnen JT, 't Hoen PA.
Source
Center for Human and Clinical Genetics, Leiden University Medical Center, Einthovenweg 20, 2333 ZC Leiden, The Netherlands. [email protected]
Abstract
BACKGROUND:
MicroRNAs are small non-coding RNA transcripts that regulate post-transcriptional gene expression. The millions of short sequence reads generated by next generation sequencing technologies make this technique explicitly suitable for profiling of known and novel microRNAs. A modification to the small-RNA expression kit (SREK, Ambion) library preparation method for the SOLiD sequencing platform is described to generate microRNA sequencing libraries that are compatible with the Illumina Genome Analyzer.
RESULTS:
High quality sequencing libraries can successfully be prepared from as little as 100 ng small RNA enriched RNA. An easy to use perl-based analysis pipeline called E-miR was developed to handle the sequencing data in several automated steps including data format conversion, 3' adapter removal, genome alignment and annotation to non-coding RNA transcripts. The sample preparation and E-miR pipeline were used to identify 37 cardiac enriched microRNAs in stage 16 chicken embryos. Isomir expression profiles between the heart and embryo were highly correlated for all miRNAs suggesting that tissue or cell specific miRNA modifications do not occur.
CONCLUSIONS:
In conclusion, our alternative sample preparation method can successfully be applied to generate high quality miRNA sequencing libraries for the Illumina genome analyzer.
PMID: 21171994 [PubMed - indexed for MEDLINE] PMCID: PMC3022920