Discrepancies between UCSC and RefSeq transcripts

[LinaGallego]

Hello everybody,

I have found the following confusing situation:
Human CELF2 gene contains a group of 14 UCSC transcripts, but only 4 RefSeq transcrip IDs are described for this gene. The point is that several UCSC transcripts are surprisingly assigned to the same RefSeq transcript ID, despite the fact that each UCSC transcript is characterized by a unique set of exons and splice sites, so the final product is clearly not the same between them.

I would really appreciate if anyone could help me to understand this discrepancy, since it is very important for my work to define the different contributions of CELF2 isoforms to a particular phenotype.

Here is one example of what I have just talked about, two different transcript (two completely different transcripts with the same RefSeq NM_001083591):

Human Gene CELF2 (uc010qbk.1) Description and Page Index
RefSeq Summary (NM_001083591)
Position: chr10:11,206,993-11,330,515 Size: 123,523 Total Exon Count: 8 Strand: +


Human Gene CELF2 (uc001ikk.2) Description and Page Index
RefSeq Summary (NM_001083591
Position: chr10:11,059,893-11,374,605 Size: 314,713 Total Exon Count: 14 Strand: +


Thanks a lot for any help!

[mastal]

Sorry if this adds even more confusion to your situation.

Have a look at Ensembl and see how many Ensembl transcript IDs map to this gene, and how they compare to the UCSC and RefSeq transcripts.

[mbblack]

Directly quoted from the RefSeq FAQ (http://www.ncbi.nlm.nih.gov/books/NBK50679/)

Why are some splice variants for my favorite gene missing in the RefSeq set?

RefSeq records that represent alternately spliced transcript variants are provided when there is experimental and/or published evidence in support of the full-length nature of the product. When transcript alignments (to the assembled genome) indicate that there is alternate splicing no assumption is made about the naturally-occurring combination(s) of alternate exons in the absence of full-length support. As a consequence, alternately spliced products are underrepresented in the RefSeq collection.

Requests to review additional transcript variants for a gene, and their possible representation in RefSeq, can be made using the Gene and RefSeq Feedback form. We encourage research groups to submit primary sequence data representing alternatively spliced transcripts to the INSDC.