Why Do Genbank And Refseq Have Different Coordinates For Mitochondrial Features?

Question

[Jeremy Leipzig]

Why are there two different start sites for the tRNA-Gln, for instance:

http://www.ncbi.nlm.nih.gov/nuccore/V00662

 tRNA            complement(4320..4400)
                 /product="tRNA-Gln"

http://www.ncbi.nlm.nih.gov/nuccore/NC_012920

 tRNA            complement(4329..4400)
                 /gene="TRNQ"
                 /gene_synonym="MTTQ"
                 /product="tRNA-Gln"
                 /note="NAR: 0597"
                 /anticodon=(pos:complement(4365..4367),aa:Gln)
                 /codon_recognized="CAA"
                 /db_xref="GeneID:4572"
                 /db_xref="HGNC:7495"
                 /db_xref="MIM:590030"

Answer 1

[Neilfws]

Sequences are re-annotated over time and GenBank archives all submissions, even those that are corrected later.

The first record is an old GenBank sequence (2006). The second is a newer reference sequence (2010), with a higher degree of curation. The COMMENT section states that it was derived from another GenBank record, J01415 (2008), in which the tRNA-Gln coordinates match the RefSeq record.

I'd speculate that tRNA-finding algorithms have improved over time and/or the curators found additional biological evidence which led them to change the start site by 3 codons.

[lh3]

In addition, mtDNA has indels. The length/coordinate may be different, too (though apparently not in your case). As I remember, there is an indel between GRCh37.p0 and the rCRS strain.

[Neilfws]

Just FYI, I ran J01415 through tRNAscan-SE (http://lowelab.ucsc.edu/cgi-bin/tRNAscan-SE.cgi) and Rfam (http://rfam.sanger.ac.uk/). I also looked at the pre-computed Rfam result. All three place tRNA-Gln at 4329 - 4400.