[BioC] TLR4 in illuminaHumanv3.db

[Mark Dunning]

Hi Gavin,

This probe is classified Bad based on the fact that maps to a
RepeatMasked element of the genome, specifically a LINE repeat in this
case. Previously, we have found repeat mapping probes to be unreliable
(Barbosa-Morais et al NAR 2010) and result in spurious signal. It
would be advisable to take this information into account when
interpreting the data for this probe.

The version of the annotation package you are using annotates any Bad
or No Match probes as NA. However, now is a good time to mention that
the illumina annotation packages that I maintain for Bioconductor 2.8
will have the probe quality and other useful metrics available to the
user.

Regards,

Mark

On Thu, Apr 14, 2011 at 8:46 AM, Gavin Koh <gavin.koh at gmail.com> wrote:
> Dear Dr Dunning,
>
> I am looking at Illumina Human WG6 v3 data in an infection and
> re-annotated the results using the illuminaHumanv3 package.
> After re-annotation, TLR4 disappeared. TLR4 seems to be represented by
> only one probe in that array (ILMN_1706217, sequence
> GCTGCCACATGTCAGGCCTTATGCTAAGGGTGAGTAATTCCATGGTGCAC). When I BLAT for
> that sequence, I get only one hit on chromosome 9 that falls squarely
> into the middle of an exon.
> I thought I might be using the package incorrectly, but
> mget("ILMN_1706217",illuminaHumanv3SYMBOL) returns "NA" and
> mget("ILMN_1731048",illuminaHumanv3SYMBOL) gives me "TLR1", which is
> correct.
>
> Can you help at all, please? I do not know why the probe seems to have
> been marked bad.
>
> Thank you in advance for your time,
>
> Gavin Koh
> Team 15, Wellcome Trust Sanger Institute
>
> OSX Snow Leopard
> R 2.12.1
> illuminaHumanv3.db 1.8.0
> illuminaHumanv3() reports
> illuminaHumanv3SYMBOL has 25556 mapped keys (of 48803 keys)
> DB schema: HUMANCHIP_DB
> DB schema version: 2.1
> Organism: Homo sapiens
>