Fruit fly paper retracted when gene turns out not to code for a protein as claimed

1.cover-sourceThe Journal of Insect Science is retracting a paper on the genetics of a fruit fly after discovering one of the genes the authors sequenced doesn’t appear to code for a protein.

The paper, “Molecular phylogeny and identification of the peach fruit fly, Bactrocera zonata, established in Egypt” was published in 2011, and compared sequences of the Egyptian species to those from species in other regions. It has not yet been cited, according to Thomson Scientific’s Web of Knowledge.

Phyllis Weintraub, the editor-in-chief of the journal, told us she thinks that the paper’s fatal mistake stemmed from “bad science instead of deliberate falsification.”

The retraction notice should go live on the site today, according to Lisa Junker, director of publications and communications for the Entomological Society of America, which publishes the journal. Here’s the text:

The Journal of Insect Science was notified of concerns about the nucleotide sequence of a gene reported in this paper. Upon further investigation by the Journal, it was determined that the sequence does not translate to a protein and that comparison to other COI gene isolates of Bactrocera zonata shows a lack of sequence similarity. In addition, phylogenetic analyses presented in the paper were not appropriate. Furthermore, the paper does not specify that voucher specimens were deposited, and therefore it is not possible to confirm the identification of the Egyptian specimen as B. zonata.

The journal has no evidence of intentional author misconduct in sequencing the gene that was isolated. However, based on the issues with the gene sequence and phylogenetic analysis, the Journal of Insect Science believes this paper to be sufficiently flawed that its conclusions cannot be relied upon. Therefore, the journal is retracting this paper.

Weintraub said a group of researchers contacted the journal and claimed that:

The COI (cytochrome oxidase I) sequence: a) contained many stop codons, b) did not show any similarity to any of the other COI accessions on GeneBank, c) did not code for any protein.

She continued:

On the basis of their accusations I undertook a full review of the manuscript and checked each point they raised, and passed all of this information on to the Author Misconduct Panel who verified all.

She explained:

One can easily go to GenBank and put in the authors new accession number: http://www.ncbi.nlm.nih.gov/nuccore/GQ225768 .  This page shows you all of the information about their accession by reading down the left hand column.  At the bottom you will see ORIGIN – this is the sequence of their COI.   On the right hand side near the top you will see Analyze This Sequence.  Click on Run BLAST  (Basic Local Alignment Search Tool), which will check the alignment of all COI accessions for the Peach Fruit Fly in Genebank.  Only one sequence turns up – this sequence is only related to itself.  Now, back to the first page of Genbank, put in one of the New Zealand or Indian accession numbers from the attached manuscript.  You will see that below ORIGIN is the protein from the translated sequence.  No protein is given for the Egyptian accession.

Further to point 1a, the reason that no protein was given was because the sequence they submitted contained many stop codons, in red below.  No protein could be translated from the short nucleotide sequences.  In other words it was not a functional gene that they sequenced.

GQ225768   red highlights are stop codons: TAG, TAA, TGA 

tcnaggtttn ggaaccttgt cttttatttt cggagcctgg agcaggtata agttgggaaa

catctcttaa gcaatttnca gttctgtgct cgagcttagg gaccacccca ggtagctact

aaattgggta gaatgatcca ggattgttat gaatggtaat tggtgaacag cacacagctt

tccgctaaat gaatttttcg tttttataag taatgaccct attataattg gggaaggggt

tttgagaaaa attggacttt gttcccccct taaatgattt aagggaagcc accccgacga

tagccatttc cccacgaatg gaataatata agattttgat tattgacctc ctttcccctt

tacgcttga

The notice said there was “no evidence of intentional author misconduct in sequencing the gene that was isolated.” Weintraub said the findings were “bad science instead of deliberate falsification:”

…[T]hey sequenced something, but certainly not COI, and continued all of their results based on that sequence.  The falsification was in claiming that it was similar to the New Zealand and Indian sequences.  I guess they thought no one would ever actually check those sequences or go back to GenBank.  We felt that if they had deliberately falsified the sequence, they probably would have presented a sequence very similar to the New Zealand sequences – a base pair or two different, for example – and then continue with their Results of the bootstrap and phylogenetic analysis, and all of their Discussion.  This they did not do, they continued with the sequence that they obtained.

She also said:

I think they made an honest mistake when they thought they were sequencing COI; however, all of their conclusions after that were not honest mistakes.

Before the retraction was made final, the authors of the paper were given 30 days to respond to inquiries from the journal, said Weintraub, but did not. We also reached out the authors, and we have not heard back. We’ll update if we do.

Weintraub said it has been a time of change at the journal:

…Journal of Insect Science strives to publish quality manuscripts that have been checked for plagiarism and other author misconduct issues.  There have been a number of transitions in the last year and a half:  The founder and EiC died and the Journal was without an EiC for about a year, the Journal was acquired by the Entomological Society of America, and all of their publications taken over by Oxford University Press.  I have been the editor-in-chief for just 6 months now and have already dealt with a number of author misconduct issues.  I am striving to improve the journal.  However, the reviewers and Subject Editors are human and mistakes happen, such as we see with this manuscript.

Like Retraction Watch? Consider supporting our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post. Click here to review our Comments Policy.

3 thoughts on “Fruit fly paper retracted when gene turns out not to code for a protein as claimed”

  1. A question for Dr. Weintraub. If the paper was indeed “bad science”, then why did the peer review of her journal not pick this up during peer review?

    1. I am actually impressed with the thorough manner that the present EiC has handled this case. The peer review system is not perfect and the probability of errors slipping through is not small – for any journal/discipline. So, I would not bother to waste effort on asking questions about the peer review. I hope the present EiC will follow up on “other misconduct issues as well”, as she implies in her response.

  2. Anonymous
    A question for Dr. Weintraub. If the paper was indeed “bad science”, then why did the peer review of her journal not pick this up during peer review?

    Well, nobody checked the sequence, nobody did any alignments and the authors don’t actually show any alignment. So reviewers and editors just believed what the authors wrote. Happens all the time….

    I wonder where they got the sequences for their primers? And is it normal to do mitochondrial gene analysis on RNA/cDNA? Why not simply use DNA?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.