How to determine the primary Uniprot accession number from a set of accession numbers?

Question

Given a list of Uniprot IDs that are linked to an Ensembl gene ID, is there a way to systematically determine the UniProt accession corresponding to the protein product of the gene's primary transcript with no other information?

According to ExPasy

Researchers who wish to cite entries in their publications should always cite the first accession number. This is commonly referred to as the 'primary accession number'. 'Secondary accession numbers' are sorted alphanumerically.

But what if the order has been jumbled or compiled from a different source and resorted?

For example:

Ensembl:

ENSMUSG00000035642

Uniprot:

Q8R0P4, Q8CF11, D6RJK8, D6RJJ4, D3Z442, D3Z1Q3, D3YZD8, D3YY39, D3YX09, D3YWY5

This question is cross listed on the Biology stack-exchange.

Answer 1

Reposting my answer from the related ticket in the Biology section:

I think there is an issue with the terminology. The "primary" accession number, is the first accession number in cases where an entry has more than one accession number, as described in http://www.uniprot.org/help/accession_numbers:

Entries can have more than one accession number. This can be due to two distinct mechanisms:

a) When two or more entries are merged, the accession numbers from all entries are kept. The first accession number is referred to as the

‘Primary (citable) accession number’, while the others are referred to as ‘Secondary accession numbers’. These are listed in alphanumerical order.

b) If an existing entry is split into two or more entries (‘demerged’), new ‘primary’ accession numbers are attributed to all

the split entries while all original accession numbers are retained as ‘secondary’ accession numbers.

Example: P29358 which has been ‘demerged’ into P68250 and P68251.

Both reviewed and unreviewed entries can have primary accession numbers.

What you probably mean, as previous posters understood, are accession numbers of reviewed entries as opposed to unreviewed ones.

In that case, you can indeed add "reviewed:yes" to your query, e.g. when you are using the UniProt ID mapping, http://www.uniprot.org/help/uploadlists

Answer 2

Consider the circumstances in which entries will have more than one accession (taken from the same page you linked in the original post).

Entries will have more than one accession number if they have been merged or split. For example, when two entries are merged into one, the accession numbers from both entries are stored in the AC line(s).

If an existing entry is split into two or more entries (a rare occurrence), the original accession numbers are retained in all the derived entries and a new primary accession number is added to all the entries.

As mentioned in the comments, this does not imply that there is or should be a single "primary" UniProt accession associated with (for example) an Ensembl gene ID. In some cases, a single protein record may have multiple accessions due to annotation changes, and the suggestion from your original post is to only report the primary accession. However, there may be multiple proteins associated with an Ensembl gene ID, and each protein will have its own accession (or set of accessions).

Choosing a single representative isoform for a gene is an entirely different question.

Answer 3

Not with no other information, no. There is no way of looking at a list of accessions and guessing which one is primary. Primary accessions are just accessions like any other; they don't have any specific characteristics.

This is further complicated by the fact that each gene can (and usually does, for higher eukaryotes) have multiple transcripts, each of which can produce a different protein product. So in order to go from gene identifier to a protein, you need to first find the relevant transcript.

So, as a first step, get the primary transcript IDs for each of your genes and then retrieve the protein accessions associated to them.