I am trying to figure out why thousands of transcript clusters are not mapped in ragene20sttranscriptcluster.db while they are perfectly present in the manufacturer's materials. Just one example: transcriptcluster 17610996 is well described in the manufacturer's annotation (support files at https://www.thermofisher.com/order/catalog/product/902124), with chromosome coordinates given (chr1+:23093113-23164282), gene symbol (Enpp3), transcript ID (NM_019370) etc. But in ragene20sttranscriptcluster.db this cluster is not mapped in any of the sets besides ragene20sttranscriptclusterACCNUM - at what stage was it 'lost'?

And what would you recommend as a solution to map those missing transcripts? (in my case, I need to get gene symbol for each transcript)

library("ragene20sttranscriptcluster.db")

# example 1: transcript is found and mapped, all OK
'17736437' %in% mappedkeys(ragene20sttranscriptclusterCHR) # TRUE

# example 2: transcript is mapped only in ragene20sttranscriptclusterACCNUM
'17610996' %in% mappedkeys(ragene20sttranscriptclusterACCNUM) # TRUE
'17610996' %in% mappedkeys(ragene20sttranscriptclusterCHR) # FALSE
'17610996' %in%       keys(ragene20sttranscriptclusterCHR) # TRUE

# example 3: transcript is not mapped even in ragene20sttranscriptclusterACCNUM
'17707702' %in% mappedkeys(ragene20sttranscriptclusterACCNUM) # FALSE
'17707702' %in% mappedkeys(ragene20sttranscriptclusterCHR) # FALSE
'17707702' %in%       keys(ragene20sttranscriptclusterCHR) # TRUE

The answer is a bit more complicated than what you might think. First, using the old BiMap interface (e.g., using things like ragene20sttranscriptclusterCHR directly) is probably suboptimal at this point, and you should be using either select or mapIds instead.

Second, these days the ChipDb objects only contain one table that maps the chip ID to an Entrez Gene ID, and all of the heavy lifting is done by (in this case) the org.Rn.eg.db package. That's a convoluted way of saying that the chr/pos information that Affy (or more correctly ThermoFisherAllScientificCompaniesInOne) supply are not used at all. What you get from the annotation csv from Affy is the actual position for the set of probes, whereas what you would get from the ChipDb would be the position of the gene being measured:

> select(ragene20sttranscriptcluster.db, "17610996", c("CHR", "CHRLOC","CHRLOCEND"))
'select()' returned 1:many mapping between keys and columns
   PROBEID CHR   CHRLOC CHRLOCCHR CHRLOCEND
1 17610996   1 21613147         1  21684482
2 17610996   1       NA      <NA>        NA
Warning message:
In .deprecatedColsMessage() :
  Accessing gene location information via 'CHR','CHRLOC','CHRLOCEND' is
  deprecated. Please use a range based accessor like genes(), or select()
  with columns values like TXCHROM and TXSTART on a TxDb or OrganismDb
  object instead.

And third, as you can see from the warning, we have split the annotation data (things like what the gene is, what it does, the protein name, etc) off from the positional information (where the gene is in the genome, where its exons are , the CDS, etc). People in general don't particularly care where the probes for an affy probeset align, but instead want to know where the gene it measures comes from.

But in the end you seem to just want the symbol and/or transcript ID:

> select(ragene20sttranscriptcluster.db, "17610996", "SYMBOL")
'select()' returned 1:many mapping between keys and columns
   PROBEID       SYMBOL
1 17610996        Enpp3
2 17610996 LOC103690935
> select(ragene20sttranscriptcluster.db, "17610996", "ACCNUM")
'select()' returned 1:many mapping between keys and columns
    PROBEID       ACCNUM
1  17610996     AAB61535
2  17610996     AAB61536
3  17610996     AAH97326
4  17610996     AJ250374
5  17610996     BAA06333
6  17610996     BC097326
7  17610996     CAA88029
8  17610996     CAB59427
9  17610996       D30649
10 17610996     EDL87774
11 17610996     FQ218696
12 17610996    NM_019370
13 17610996    NP_062243
14 17610996       P97675
15 17610996       U78787
16 17610996       U78788
17 17610996       Z47987
18 17610996 XM_008758654
19 17610996 XP_008756876