Hi Sean, Thanks for your reply. I should have been more specific. I am mapping to the Zv9 cDNA build that I have downloaded from Ensembl. Best, Ravi On Feb 1, 2011, at 12:36 PM, Sean Davis wrote: > > > On Tue, Feb 1, 2011 at 12:19 PM, Ravi Karra <ravi.karra@gmail.com> wrote: > Hello all, > > Thank you for maintaining this very informative mailing list. I am currently working with some gene expression data using the Nimblegen Zebrafish 12 x135K Expression platform. Unfortunately, my analysis has been limited by annotation. The probes were derived from a variety of sources, many of which are now out of date. > > A bit reluctantly, I would like to remap the array to the latest zebrafish genome available (Zv9) at Ensembl so that I can then use biomart for subsequent annotation. I am able to map most of the probe sequences in the .ndf file to the Zv9 using matchPDict (basically using a minimally modified version of the code at http://manuals.bioinformatics.ucr.edu/home/ht-seq#TOC-Mapping-of-Affy- Probes-to-Transcrip). > > Mapping expression probes to the genome is perhaps not the best way to go. You should map, instead, to the transcripts. This is not so big a deal for 25mers (and so the vignette using affy is probably not unjustifiable). However, if you are using Nimblegen 50-60mers, then you ought to consider mapping to the transcripts. If a probe does not map to a transcript, you can probably just ignore it. An alternative is to be more exhaustive and map to things like ESTs, etc. > > > Once I do this, I am stuck. How can I use these results to make an annotation file with PDInfoBuilder? How do I handle probes that do not align with Zv9? Can I get still use rma? > > > Some of the probes may not align due to the probes crossing intron- exon boundaries. These are highly relevant probes--hence my note above. > > I'll leave the actual annotation details for others on the list. Sorry, I know that was the main point of your original question. > > Sean > [[alternative HTML version deleted]]