Connectivity Map - Unable To Load Probe Ids

Question

[Olivier]

Hello all

I've been unsuccessfully trying to load up- and down-regulated genes in CMAP (http://www.broadinstitute.org/cmap/signature). Here's what I get (same thing for all probe ids..):
invalid tag files:
down tags: probe '31850_at' invalid, not found in database
down tags: probe '40805_at' invalid, not found in database
down tags: probe '1269_at' invalid, not found in database
down tags: probe '40513_at' invalid, not found in database
down tags: probe '32290_at' invalid, not found in database

Here's a sampe from the signature, if someone wants to try:
Up-regulated probesets
36025_at
38646_s_at
35909_at
32640_at
39946_at
34622_at
37674_at
37637_at

Down-regulated probesets
39232_at
35494_at
40230_at
36215_at
36508_at
36391_at
36363_at
39943_at
32693_at
33460_at
33610_at
33611_g_at
37170_at

Grateful if someone could help. Could it be possible that all probe ids have changed?? I got these from AILUN, using my gene lists and a taxid of 9606 (human).

It's part of a work in progress (Bioinformatics study group) to reproduce work from the paper "Computational repositioning of the anticonvulsant topiramate for inflammatory bowel disease"

Answer 1

[David Westergaard]

As I recall, CMAP stricly required all probesets to come from the Affymetrix GeneChip Human Genome U133A Array. Anyone else will result in a failure. It took me quite a few hours to figure out when I had to make use of CMAP.

[Olivier]

In the above-mentioned paper (published in 2011), that's what they used: Dataset GDS2642. I fetched it from ncbi and after a quick check that's what I found as annotation platform: Affymetrix Human Genome U95 Version 2 Array Is there something I'm missing somewhere?

Thanks a lot for the tip.

[David Westergaard]

No, you're not. You just have to reformat it before using it as input to CMAP.

So basicly you have to do the transform:

Affymetrix Human Genome U95 -> Gene identifier -> Affymetrix Human Genome U133

[Olivier]

It works now! Thanks a lot David. By the way, how did you know about the annotation platforms - I couldn't find that info on cmap? Got a lot of duplicate mappings. I assume I can pick any?

If anyone's interested about the affy id conversion, here's the link: http://central.biomart.org/converter/#!/ID_converter/gene_ensembl_config_2

[David Westergaard]

The information is sort of there, hidden between the lines. If you go to help and go under "f" for feature_set, it's sort of written that CMAP uses that array.

I had the issue with duplicate mappings as well. I'm not really sure what the best solution is. You should probably try to find the probe that either (a) maps to the same genes on whatever array you're converting from, or (b) the one that has the biggest intersect with the array you're trying to convert from.