Those are from two different strains, so it would not necessarily be expected that the sequences would be the same regardless of the fact that the species is the same in both cases.

In fact, a smith-waterman alignment of those two accessions actually does give an alignment of 99.9% similarity over the 1471bp of AY584477, so they very clearly are in fact the same, within any reasonable expectation for 16s rRNA between two strains of the same bacterial species (I just threw them through EMBOSS WATER at http://www.ebi.ac.uk/services), see the following output summary:

#=======================================

# Aligned_sequences: 2

# 1: AY584477

# 2: NR_102798

# Matrix: EDNAFULL

# Gap_penalty: 10.0

# Extend_penalty: 0.5

#

# Length: 1471

# Identity: 1469/1471 (99.9%)

# Similarity: 1469/1471 (99.9%)

# Gaps: 0/1471 ( 0.0%)

# Score: 7337.0

#=======================================

Were you expecting them to literally be 100.00% identical? If so, why would you? Any two random individuals from the same species (ignoring strains) might differ by at least as much as these two, even for a highly conserved gene. Aside from being from different strains, these two sequenced regions are not identical in length, were done years apart, and were created with different methods.

As for the differences in length, the former is a refseq (curated) entry from a whole genome sequencing project. The latter is a GenBank (un-curated) entry from a PCR-based phylogenetic study. Usually, any PCR based gene sequencing will include a sub-fragment of the whole gene, unless the primers are well displaced from the boundaries of the gene of interest. If the primers are close to or inside of either end of the gene, then the final, high confidence sequence product will inherently be less than the whole gene (especially if the PCR amplification primers are also used as the sequencing primers).

Dear Dr. Michael Black,

Thank you for you information.

From some years, I obtained some sequences of the clones of the V3 16S rRNA genes. Using Chromas and Blast to determine the closest known relative species, One of these sequences had an Accession no. (gb|AY584477.1) for Streptococcus parauberis . when I tried to do the same sequence, from some days, I obtained (NR_102798.1) as Accession no for the same species.

My question now is :Can I use weather the new accession no (NR_102798.1) or the old one (gb|AY584477.1) in an article? or I must use the new one.

Best regards,

Gaber El-Baradei

The fundamental difference between an NR_12345 accession and a Genbank (gb|AY12345) accession is that the NR_ number is from RefSeq, and RefSeq entries have undergone some curation, and will be updated if errors are found. The Genbank sequence is exactly as was submitted by the authors; getting a Genbank sequence corrected is more difficult, and it has not undergone the curation processes used for RefSeq. In general, RefSeq sequences are of higher quality, so it would be better to use the RefSeq accession.

Thanks for your information,

now I have some sequences in a fasta form, when I use BLAST, I always obtain a (NR_ number). Is there a way to get a gb-number depending on Genbank sequence submitted by the authors not on RefSeq?

Thanks again