I want to get Entrez IDs for Affymetrix probe sets (hgu133a) to map them on genome-scale metabolic models (GSMMs). Generally, GSMMs use Entrez gene IDs; therefore, to integrate gene expression profiles with these reconstructions, ID conversion between probe sets in respective Affy platform and Entrez IDs is required. The problem with this conversion in Bioconductor is the presence of multiple mapping between identifiers. A simple example would be:
> select(hgu133a.db, c("200080_s_at"), c("SYMBOL","ENTREZID", "GENENAME"))
'select()' returned 1:many mapping between keys and columns
PROBEID SYMBOL ENTREZID GENENAME
1 200080_s_at H3F3A 3020 H3 histone, family 3A
2 200080_s_at H3F3B 3021 H3 histone, family 3B (H3.3B)
3 200080_s_at H3F3AP4 440926 H3 histone, family 3A, pseudogene 4
The question is, what is the choice here? 3020, 3021 or 440926?
One should notice that the resulting Entrez IDs will be used for GPR (gene-protein-reaction) purpose; therefore, the expression level of all these three Entrez IDs is the same.
Thanks in advance for sharing your thoughts.
Hi Oveis,
If it doesn't matter which one of the entrez you use for multiple mapping, you could use data
Dear Austin,
Thanks for your answer. There are 2 problems with GPR: 1-The presence of redundant IDs. For instance, if we have a expression profile containing 200 probe sets, its conversion to Entrez may be a vector with size of 300 (multiple mapping); and 2-Several probe sets may map to the same Entrez ID.
Removing duplicates could be problematic, since some Entrez IDs (and as the results, some expression levels) could be overlooked. It seems to me that, one could use simply the average of various expression levels when several identical Enrez IDs appear in a expression profile. Nevertheless, the biological implication of this approach is not clear to me.
Hi Oveis,
If you plan to average the levels, I would use geometric mean to remove outlier influence. However, since it seems like you are using the "s" set of affy probes (I did not see that when I first responded), they cross-hybridize to the same sequences as the design and it would be more difficult to determine biological impact as it is an inaccurate "group, probe, then analyze" method just using the R packages.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics