To my knowledge, the current Illumina annotation packages are based on Illumina annotation's TargetID column which means they are created for the gene profile data. In gene profile data the possibly several different probe types for the same TargetID have been averaged. However, there are usually only couple of different probe types for a TargetID. This means that averaging is not very reliable and this is why I prefer to analyse the different probe types separately. I have also seen cases where one of the probe types for a TargetID is showing high expression values and the other is showing very low expression values. In this case I would still be interested in the information that one of the probe types has, even the other one does not seem to work. If I would use the gene profile, the information for that interesting probe type would be most likely lost. The problem is that when using probe profile data, it is not possible to combine it with Bioconductor's Illumina annotation packages since they only support TargetIDs and not ProbeIds. So my question is: has there been any consideration or would it be possible to start creating Bioconductor annotation packages based on Illumina's ProbeIds as well? I happen to know how to create similar kinds of packages by myself (with AnnBuilder), but for the sake of many other users I think these ProbeId based packages should be made publicly available. Thanks, Asta Laiho

Hello Asta, I have made Illumina annotation packages based on probe IDs. They are not a part of Bioconductor but I can send them to you if you can receive large emails. I may also be able to set up a temporary ftp site. Let me know if you are interested. Lynn Amon Asta Laiho wrote: > To my knowledge, the current Illumina annotation packages are based on Illumina annotation's TargetID column which means they are created for the gene profile data. In gene profile data the possibly several different probe types for the same TargetID have been averaged. However, there are usually only couple of different probe types for a TargetID. This means that averaging is not very reliable and this is why I prefer to analyse the different probe types separately. I have also seen cases where one of the probe types for a TargetID is showing high expression values and the other is showing very low expression values. In this case I would still be interested in the information that one of the probe types has, even the other one does not seem to work. If I would use the gene profile, the information for that interesting probe type would be most likely lost. > > The problem is that when using probe profile data, it is not possible to combine it with Bioconductor's Illumina annotation packages since they only support TargetIDs and not ProbeIds. So my question is: has there been any consideration or would it be possible to start creating Bioconductor annotation packages based on Illumina's ProbeIds as well? I happen to know how to create similar kinds of packages by myself (with AnnBuilder), but for the sake of many other users I think these ProbeId based packages should be made publicly available. > > Thanks, > Asta Laiho > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor >

Asta Laiho wrote: > To my knowledge, the current Illumina annotation packages are based on Illumina annotation's TargetID column which means they are created for the gene profile data. In gene profile data the possibly several different probe types for the same TargetID have been averaged. However, there are usually only couple of different probe types for a TargetID. This means that averaging is not very reliable and this is why I prefer to analyse the different probe types separately. I have also seen cases where one of the probe types for a TargetID is showing high expression values and the other is showing very low expression values. In this case I would still be interested in the information that one of the probe types has, even the other one does not seem to work. If I would use the gene profile, the information for that interesting probe type would be most likely lost. > > The problem is that when using probe profile data, it is not possible to combine it with Bioconductor's Illumina annotation packages since they only support TargetIDs and not ProbeIds. So my question is: has there been any consideration or would it be possible to start creating Bioconductor annotation packages based on Illumina's ProbeIds as well? I happen to know how to create similar kinds of packages by myself (with AnnBuilder), but for the sake of many other users I think these ProbeId based packages should be made publicly available. One possibility is to use the lumi package and their annotation packages. See the discussion in their vignette about nuIDs, which are representations of the actual sequence of a probe. Perhaps Simon Lin or Pan Du will comment in more detail. Sean