Hi James On Mon, Feb 28, 2011 at 17:42, James W. MacDonald <jmacdon@med.umich.edu>wrote: > Hi Assa, > > > On 2/28/2011 4:30 AM, Assa Yeroslaviz wrote: > >> Hi BioC users, >> >> I'm working on a set of affy chips for miRNA. I would like to run some >> quality control tests but none of the functions I am trying seems to work. >> I have tried the AffyQCReport, affycoretools and simpleaffy. But every >> time >> I am getting the error message that my qcdef file is not there. >> I downloaded the files from affymetrix and rename the miRNA- 1_0_2Xgain.CDF >> into mirna102xgaincdf. >> > > Not sure what you mean here. Do you in fact mean that you used biocLite() > to get the correct cdf package for this chip? Which is what you should be > saying. There is almost never a reason for you to be getting the 'raw' cdf > from Affy. Yes you are right. It didn't help much renaming the file, so I tried to do it according to the instructions in the simpleaffy package. Unfortunately it is not exactly easy to follow them. It doesn't say much about how to create a qcdef file. I download from affymetrix the information I could find and tried to create a mirna102xgaincdf.qcdef fie. This is what I came up with: array mirna102xgaincdf alpha1 0.05 alpha2 0.065 spk bioB AFFX-BioB-3_at spk bioC AFFX-BioC-3_at spk bioD AFFX-BioDn-3_at spk creX AFFX-CreX-3_at I didn't change the alpha values as I don't know what they exactly means. The spike-ins are the same as in other arrays. My problem begins with the ratios. There are no house keeping genes on the array, so I don't know what to take. This is what Affymetrix has to say about the control probes on these arrays: Control target content 95 GC-binned background probe sets ≤ 94 probes/set 54 BioB, C, D, Cre r2 hybridization control probe sets 11 probes/set 9 BioB, C, D, Cre hybridization control probe sets 20 probes/set 22 oligonucleotide spike-in control probe sets 10 identical probes/set (i.e., each probe is tiled 10 times) 10 identical probe sets for human 5.8s rRNA(gi555853) 11 probes/set But the qcdef file can have only spk for spik-ins and ratio for ratios of gene pairs. I am not exactly sure, whether it makes sense to add the cre, bioc,d,b as ratios to the qcdef file, something like that: AFFX-r2-Ec-c1-bioB-3_at/AFFX-r2-Ec-c1-bioB-5_at AFFX-r2-Ec-c1-bioC-3_at/AFFX-r2-Ec-c1-bioC-5_at AFFX-r2-Ec-c1-bioD-3_at/AFFX-r2-Ec-c1-bioD-5_at I will try this one and will tell what happens The array has extra probes with the names spike_in 405 spike_in-control-17_st Unknown 406 spike_in-control-1_st Unknown 407 spike_in-control-21_st Unknown 408 spike_in-control-23_st Unknown 409 spike_in-control-27_st Unknown 410 spike_in-control-28_st Unknown 411 spike_in-control-29_st Unknown ... 426 spike_in-control-8_st Unknown Can it be that these suppose to be the spike ins for the array and the cre, biob-d are instead of the house keeping genes? > > than I did >> >>> setQCEnvironment("mirna102xgaincdf", path=getwd()) >>> >> >> which apperently works just fine, but: >> >>> QCReport(rawData) >>> >> Error in setQCEnvironment(cdfn) : >> Could not find array definition file ' mirna102xgaincdf.qcdef '. >> Simpleaffy does not know the QC parameters for this array type. >> See the package vignette for details about how to specify QC parameters >> manually. >> > > Right. Which means you should read the package vignette, where they give > some indication what to do to specify QC parameters. Did you? > > This does raise a bit of a problem, as I don't see any actin or GAPDH > probes on this chip. Perhaps you can get away with just specifying something > for the spike-in controls. Or maybe not. You will have to try and see what > happens. yes I will try it > > > > In addition: Warning message: >> In data.row.names(row.names, rowsi, i) : >> some row.names duplicated: >> >> 3,4,8,9,10,14,15,16,20,21,22,23,28,29,30,31,35,36,37,38,43,44,45,49 ,50,51,52,56,57,58,62,63,64,68,69,70,75,76,77,81,82,83,84,88,89,90,94, 95,96,97,101,102,103,107, >> >> 108,109,113,114,115,120,121,122,126,127,128,132,133,134,135,139,140 ,141,145,146,147,152,153,154,158,159,160,164,165,166,167,172,173,174,1 96,197,198,199,200,201,207,208,209,210,211,217, >> >> 218,219,220,221,227,228,229,230,231,237,238,239,240,241,245,246,247 ,253,254,255,256,257,258,264,265,266,267,268,274,275,276,277,278,284,2 85,286,287,288,294,295,296,297,298,304,305,306, >> >> 307,308,314,315,316,317,318,324,325,326,327,328,335,336,337,338,339 ,345,346,347,348,349,355,356,357,358,359,365,366,367,368,369,375,376,3 77,378,379,385,386,387,388,389,396,397,398,399, >> >> 400,404,405,406,407,413,414,415,416,417,424,425,426,427,428,434,435 ,436,437,438,444,445,446,447,448,454,455,456,457,458,464,465,466,467,4 68,474,475,476,477,478,485,486,487,488,489,495, >> >> 496,497,498,499,505,506,507,508,509,515,516,517,518,519,525,526,527 ,528,529 >> [... truncated] >> >> I would like to know whether there are some specific qc tools or >> R-packages >> for the miRNA from Affymetrix (the older version of the arrays). >> > > I don't know of any miRNA-specific qc packages. But many of the functions > in simpleaffy, AffyQCReport, and arrayQualityMetrics should work. However, > you might need to pick and choose what to run rather than simply running the > main wrapper function, as many of the functions are likely not to work for > this chip. > Unfortunately they ask for a qcdef file. > > Best, > > Jim > > > >> I did an MA plot of the arrays and they looks fine to me, but I would >> still >> like to sun some more tests. >> >> THX, >> Assa >> >> [[alternative HTML version deleted]] >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor@r-project.org >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor >> > > -- > James W. MacDonald, M.S. > Biostatistician > Douglas Lab > University of Michigan > Department of Human Genetics > 5912 Buhl > 1241 E. Catherine St. > Ann Arbor MI 48109-5618 > <734-615-7826>734-615-7826 > ********************************************************** > Electronic Mail is not secure, may not be read every day, and should not be > used for urgent or sensitive issues > [[alternative HTML version deleted]]