Hold the presses: maybe tardigrades don’t have so much horizontal gene transfer after all

December 4, 2015 • 10:30 am

About a week ago I discussed a new paper by Boothby et al. in the Proceedings of the National Academy of Sciences (US) with a stunning finding: the sequenced DNA of the tardigrade species Hypsibius dujardini showed that about 17% of its genome comprised sequences taken from distantly-related species—mostly bacteria. This was the most pervasive example of horizontal gene transfer (HGT) known in animals, though bacteria often have such high levels of HGT.

But maybe that conclusion was premature—or wrong. A new manuscript by Georgios Koutsovoulos et al. (reference and free download at bottom), posted at the website bioRχiv, suggesst that Boothby et al.’s results may have been due to contamination of their DNA sample with bacteria, and the level of HGT in this species may be much lower.

Do be aware, when evaluating the discrepancy between the two papers, that the Koutsovoulos et al. manuscript hasn’t been refereed: the authors simply posted a manuscript—presumably submitted to some unspecified journal—on a public website. But let’s compare them anyway. 

Although Boothby et al. argued that they cleaned up their sample to eliminate contamination, Koutovoulos seem to have taken more stringent precautions. And the sequencing of their sample, compared to the published sequences of Boothby et al., showed a very different result. First, about 30% of Boothby’s DNA appears to represent contaminant material.

Granted, their sample was a different isolate of the species (started from one asexually-reproducing female), but their culture and that of Boothby et al. were separated by only fifteen years of divergence, so it’s unlikely that the differences in DNA content result from different genetic constitution of the isolates. 

How much of the tardigrade genome is still composed of foreign DNA in the new species? We don’t know completely, but Koutsovolos et al. report that they looked at 23,021 protein-coding genes in the H. dujardini genome, and found only 36 genes that appeared to have a bacterial origin. That’s a proportion of only 0.16%—a far cry from 17%.

Koutsovoulos et al. also note that H. duardini is not known to undergo the drying and rehydration process known as cryptobiosis—a process that takes in lots of water from the environment, destabilizes membranes, and breaks DNA: an idea way to take foreign DNA into your genome. But if this tardigrade doesn’t do that—and I don’t recall Boothby et al. mentioning this fact, though they might have—then it makes it even less likely that the 17% HGT figure is correct.

As I said, the “replication” study hasn’t yet been published after peer review, so we’ll have to suspend judgment. Resolution of this problem will await the publication of the Koutsovoulos et al. paper, and, ideally, a third study (perhaps by Boothby et al.) with very stringent precautions. After all, the press had a field day with the 17% HGT figure, which truly was astounding; and, as Hitchens said, extraordinary claims demand extraordinary evidence.

And a postscript: the methods of both papers are above my pay grade, so read both if you’re in the field and want to judge the results.


15 thoughts on “Hold the presses: maybe tardigrades don’t have so much horizontal gene transfer after all

  1. This should be easy to verify studying the genome sequence contigs. If it is contamination the sequences should not appear in contigs with tardigrade sequence

    1. According to the authors they did PCR analysis to confirm physical linkage of sequences attributed to originate from HGT and metazoan (tardigrade)sequences; they also used PacBio for confirmation. From the PNAS paper:

      “assembly artifacts might result in contaminant sequences assembling next to tardigrade genes. Therefore, we looked for signs of contamination and tested directly the accuracy of assembly for a sample of genes by two methods. First, we performed PCR to test
      physical linkage of pairs of genes that were found on the same genomic scaffolds….For 104 of 107 enes, we recovered amplified products of the appropriate size, confirming that for-
      eign genes are physically present within the
      H. dujardini genome”

      “Second, we used low-depth PacBio single molecule real-time (SMRT) sequencing for comparison with our initial assembly…Because each long read comes from a single DNA molecule,the resulting contigs cannot be assembly artifacts; therefore, these
      data allowed us to test whether our initial assembly produced chimeric contigs. Our two assemblies were highly congruent”

      Those controls seem pretty convincing. I guess we will have to wait until we get confirmation from a third lab!!

    2. This should be easy to verify [by (?)] studying the genome sequence contigs

      Could you translate that into English, at a gravel inspector’s pay grade?
      What I’m picking up is that you think that the authors could, relatively easily, determine which genes are near to each other (“contiguous”, in the limit), and that contaminants wouldn’t have such repeatable patterns of contiguousness?

  2. Interesting paper and thanks for the free download link. One comment though. I believe the phrase “extraordinary claims demand extraordinary evidence” was made by Carl Sagan, not Christopher Hitchens.

    1. Actually, several figures made similar statements before Sagan. The most similar might be Marcello Truzzi: “An extraordinary claim requires extraordinary proof.”

    2. Wasn’t it Hitch who said, “that which can be asserted without evidence can be dismissed without evidence?” Possibly why Jerry connected the Sagan quote with Hitch…

  3. A suggested control:
    Split your tardigrade isolate into two populations. Feed each on just one type of foodstuff (bacteria). Then sequence both populations and ignore any sequence found in one but not the other.

  4. Does anyone remember (among the older ‘fans’, of course) a study from 20 or 30 years ago claiming that tardigrades incorporate foreign DNA into their genome? I have a feeling that reports claiming the phenomenon are old. Perhaps these recent papers, which I haven’t looked at, give references.

Leave a Reply