exonLevel: Conversion of Parameter exonlevel to Integer

Description Usage Arguments Details Value Note Author(s) See Also Examples

View source: R/utils.R


Conversion of parameter exonlevel to an integer vector.


exonLevel(exonlevel = "", chiptype = "GeneChip", as.sum = TRUE)



exon annotation level determining which probes should be used for summarization; exon/genome arrays only.


chip tpye, one of ‘GeneChip’, ‘GenomeChip’, ‘ExonChip’.


logical, if TRUE an integer vector of size three will be returned, if FALSE then the levels will be split into the basic integer representations.


Conversion of parameter exonlevel to an integer; this function is a utility function, which is usually only used internally.

Following exonlevel annotations are valid for exon arrays:

core: (=8192+1024) probesets supported by RefSeq and full-length GenBank transcripts.
metacore: (=8192) core meta-probesets.
extended: (=4096+512) probesets with other cDNA support.
metaextended: (=4096) extended meta-probesets.
full: (=2048+256) probesets supported by gene predictions only.
metafull: (=2048) full meta-probesets.
ambiguous: (=128) probesets that fall within multiple genes.
affx: (=60) standard AFFX controls.
all: (=16316) combination of above (including affx).

Following exonlevel annotations are valid for whole genome arrays:

core: (=8192+1024) probesets with category 'unique', 'similar' and 'mixed'.
metacore: (=8192) probesets with category ‘unique’ only.
affx: (=60) standard AFFX controls.
all: (=9276) combination of above (including affx).

Exon levels can also be combined, with following combinations being most useful:

exonlevel="metacore+affx": core meta-probesets plus AFFX controls
exonlevel="core+extended": probesets with cDNA support
exonlevel="core+extended+full": supported plus predicted probesets

Exon level annotations are described in the Affymetrix whitepaper exon_probeset_trans_clust_whitepaper.pdf:
“Exon Probeset Annotations and Transcript Cluster Groupings”.

Parameter exonlevel determines not only which probes are used for medianpolish, but also the probes used for background calculation and for quantile normalization. If you want to use seperate probes for background calculation, quantile normalization and medianpolish summarization, you can pass a numeric vector containing three integer values corresponding to the respective exonlevel. These integers must be the sum of the integers shown above, e.g. you can use exonlevel=c(16316,8252,8252), where 8252=8192+60 for "metacore+affx".


an integer vector.


The following exonlevels are unsupported:

control->bgp->genomic: (=32768) genomic background probes.
control->bgp->antigenomic: (=65536) antigenomic background probes.
normgene->intron: (=131072) intronic controls.
normgene->exon: (=262144) exronic controls.
rescue->FLmRNA->unmapped: (=524288) unmapped mRNAs.

For whole genome arrays it is possible (but not recommended) to use all probesets by using exonlevel=c(992316,992316,992316).
For exon arrays it is possible to use e.g. exonlevel=c(1032124,1032124,631868).
However, please note that these settings are not recommended and not supported.


Christian Stratowa

See Also

rma, mas5


exonLevel("core", "GenomeChip")
exonLevel("all", "GenomeChip")
exonLevel("core+extended+full", "ExonChip")
exonLevel("core+extended+full", "ExonChip", as.sum=FALSE)
exonLevel(c(16316,8252,8252), "ExonChip")

xps documentation built on Nov. 1, 2018, 2:29 a.m.