Question: Humanht-12 V3 Expression Beadchip: How To Calculate Expression Of A Gene Having More Than One Mapped Probes?
0
gravatar for Ttnguyen
7.1 years ago by
Ttnguyen70
Ttnguyen70 wrote:

Anyone knows how? Should I take the average of all probes mapped to that gene?

gene • 2.1k views
ADD COMMENTlink modified 7.0 years ago by rowland.mosbergen0 • written 7.1 years ago by Ttnguyen70
2
gravatar for Neilfws
7.1 years ago by
Neilfws48k
Sydney, Australia
Neilfws48k wrote:

For other platforms, gene-level summaries generally use the median of probes, not the mean. You might also use a procedure called median polish.

ADD COMMENTlink written 7.1 years ago by Neilfws48k
1
gravatar for Mark Dunning
7.1 years ago by
Mark Dunning20
Cambridge, Uk
Mark Dunning20 wrote:

Hi,

Perhaps you can clarify what you mean by probes targeting the same genes? With the Illumina technology you have multiple (~20) beads that have the same probe sequence and these are known as a 'bead-type'. Combining beads for the same bead-type is done automatically for you in GenomeStudio, or in beadarray with the summarize function.

A gene can have multiple bead-types targeting it. However, averaging bead-types that target the same gene isn't really recommended for Illumina data. When a gene has multiple bead-types targeting it, the bead-types can target different isoforms of the same gene. Some isoforms might not be expressed (or have badly-designed probe sequences), so by averaging you would be diluting the signal of those isoforms that are expressed. If you want to have the data on a per-gene basis, I tend to pick the 'best' probe for each gene using a criteria such as the probe with the highest IQR. Or for a differential expression analysis should as limma, I would fit the model to all probes (probably after some filtering to remove non-informative probes) and then map to genes afterwards.

Hope this helps,

Mark

ADD COMMENTlink written 7.1 years ago by Mark Dunning20

Hi Mark, thanks a lot for your answer. Based on your explanation, I think bead-type would be more suitable word to say than probe.

ADD REPLYlink written 7.1 years ago by Ttnguyen70
0
gravatar for rowland.mosbergen
7.0 years ago by
rowland.mosbergen0 wrote:

I have been trained to show all the probes for that gene with the gene name next to it:

eg,

http://stemformatics.org/expressions/result?graphType=box&datasetID=2000&gene=stat1&db_id=56

ADD COMMENTlink written 7.0 years ago by rowland.mosbergen0
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1400 users visited in the last hour