A million citations to European Nucleotide Archive likely to be spurious

[rdmpage]

A took a random sample of 1000 "citations" of nucleotide sequences from the corpus (attached) and I tried to resolve the using NCBI's Entrez Utilities. That service returned results for 976 accessions, of which 480 (49%) were not recognised. In other words, half of these are not accession numbers for nucleotide sequences.

There are 2,276,286 "citations" to the European Nucleotide Archive in the corpus, it is likely that half of these (> 1M) are spurious.

accessions.csv