Thank You...Steve for answering. Ok...I am reading all the pdfs very carefully and I am satisfy with your answer. But please clear my some basic question... which type of reads are selected in differential expression analysis? I understand that only unique mapped reads use for differential expression analysis but I think multimapped reads also have an important role in differential expression analysis because genome has so many duplicated/paralogous genes. if I am wrong then please tell me. regards deepika ---------- Forwarded message ---------- From: Steve Lianoglou <lianoglou.steve@gene.com> Date: Tue, Sep 24, 2013 at 12:30 PM Subject: Re: [BioC] Regarding multiple hits of same read To: deepika lakhwani <lakhwanideepika@gmail.com> Cc: "bioconductor@r-project.org list" <bioconductor@r-project.org> Hi, On Tue, Sep 24, 2013 at 4:25 AM, deepika lakhwani <lakhwanideepika@gmail.com> wrote: > Hello, > > i have been trying to find out differential expression of gene using > different R packages. I have rice illumina sequencing data (pair end) with > 100 bp. i mapped the data n rice genome using tophat now i got > accepted_hit. bam file in which details of mapping is available. > > Now i am confused because it can be possible that a single read can align > on multiple position. One way to deal with reads that align to multiple (genomic) positions is to not deal with them at all. Many people only use reads that align uniquely to the genome. > When we count the reads for differential analysis > then same read is present in two different genes. This is different than what you mention above. It is possible that: (1) One read aligns to multiple places in the genome. These reads are often called "multimapped" (multimappers, etc.) and as I mentioned above, it is rather common to ignore these and to only count reads that align to a unique position in the genome. (2) It is possible for two different genes to share the same genomic locus as each other, so even though a read maps to one position in the genome, there is more than one gene that it can be assigned to. > So i have a question that > is correct or not? Can you clarify in greater detail what you are asking "correctness" for? > and i am reading genomic features R package for counting > the reads in libraries. Can anyone explain the summarizeOverlaps function? Please read through the copious documentation made available in the GenomicRanges package: http://bioconductor.org/packages/2.12/bioc/html/GenomicRanges.html There are five PDF files available there under the "Documentation" section and all of them are worth your close attention. If you still have more specific questions after reading through those, please ask those specific ones here. A generic question like "explain the summarizeOverlaps" function isn't helpful, as it is explained in multiple places in the documentation -- if there is something specific about it that is confusing, we can help you to address that. > i read manual but what is basic function of it. So what part is unclear? You'll likely also want to read through the vignette for the parathyroidSE package: http://bioconductor.org/packages/release/data/experiment/vignettes/par athyroidSE/inst/doc/parathyroidSE.pdf It shows in great detail how to go from aligned reads to "counted" genes and exons. HTH, -steve -- Steve Lianoglou Computational Biologist Bioinformatics and Computational Biology Genentech [[alternative HTML version deleted]]