Rosacea – What Causes It? News story overplays suggested connection to skin mites

Just got done reading this: Could Bacteria in Skin Mites Help Cause Rosacea? – US News and World Report.  The article leads off with a bold statement that caught my eye

“Bacteria carried by tiny mites on the skin might be responsible for the common dermatological condition known as rosacea, researchers say.”

This caught my attention because I have been reading up on skin microbes recently and though many have suggested connections between microbes and rosacea as far as I know nobody has shown any causal relationship.  And causation vs. correlation has been on my mind a lot recently.

So I read further and found some suggestive but inconclusive statements that were linked together

  • there are more of these mites on the skin of patients with rosacea than on those without
  • a bacterium (Bacillus oleronius) has been found in the mites and in people w/ rosacea
  • this bacterium can be killed with the same antibiotics that seem to have some success in treating rosacea
  • people with rosacea have an immune reaction to compounds from this bacterium 
  • another bacterium Staphylococcus epidermis also appears in patients w/ rosacea but not patients free of rosacea

And that apparently was it … not very convincing.  Sounds like just a lot of random correlations to me.  So I decided to dig deeper.  And I went to see fi I could find the paper which alas was not linked from the news story.

I googled the journal name “Journal of Medical Microbiology” and got to the web site.  The news article had said the “review paper” had come out August 30th so I clicked on the Papers In Press link and got to the paper.  I browsed the abstract, which seemed somewhat different from the gist of the news story

Rosacea is a common dermatological condition that predominantly affects the central regions of the face. Rosacea affects up to 3% of the world’s population and a number of subtypes are recognized. Rosacea can be treated with a variety of antibiotics (e.g. tetracycline or metronidazole) yet no role for bacteria or microbes in its aetiology has been conclusively established. The density of Demodex mites in the skin of rosacea patients is higher than in controls, suggesting a possible role for these mites in the induction of this condition. In addition, Bacillus oleronius, known to be sensitive to the antibiotics used to treat rosacea, has been isolated from a Demodex mite from a patient with papulopustular rosacea and a potential role for this bacterium in the induction of rosacea has been proposed. Staphylococcus epidermidis has been isolated predominantly from the pustules of rosacea patients but not from unaffected skin and may be transported around the face by Demodex mites. These findings raise the possibility that rosacea is fundamentally a bacterial disease resulting from the over proliferation of Demodex mites living in skin damaged as a result of adverse weathering, age or the production of sebum with an altered fatty acid content. This review surveys the literature relating to the role of Demodex mites and their associated bacteria in the induction and persistence of rosacea and highlights possible therapeutic options.

And then I did what usually causes me much anguish when I am at home – I clicked on the link for the full text, thinking that I would get a paywall.  And low and behold, I got the preprint of the paper.  The paper is quite interesting in many ways with lots of details about these mites I knew nothing about.  It also has a lot of detail on these two bacterial species and why the authors think they are of interest in rosacea etiology.  But no convincing evidence of any kind is presented that there is a causal connection to these bacteria or to these mites.  I leave everyone with the last paragraph of the paper

The pathogenic role of Demodex mites, as well as B. oleronius and S. epidermidis, in the induction and persistence of rosacea remains an unresolved issue. The lack of an immunological response to Demodex mites in healthy skin raises the possibility of localized immunosuppression, facilitating the survival of the mite. Hopefully, the results of further research will bring us closer to understanding the role of microbes in the pathogenesis of rosacea and assist in the development of new and more effective therapies for the treatment of this disfiguring disease.

I agree. Unresolved.

#UCDavis neurosurgeons conducted experimental surgery w/o IRB approval

Not really sure what to say about this other than that this story should be read by many/all who are interested in medical research and/or UC Davis: 2 UC Davis neurosurgeons accused of experimental surgery are banned from human research – Investigations – The Sacramento Bee.

UPDATE 7/23.  Added a summary: Two UC Davis neurosurgeons were treating terminally ill brain cancer patients with an unapproved, experimental treatment that is referred to as “Probiotic Intracranial Therapy for Malignant Glioma”.  The treatment involved purposefully infecting patients brains with a bacterium Enterobacter aerogenes apparently because of prior anecdotes and case reports that suggested that patients with these brain cancers who also had brain infections might live longer than those with the cancer but without the infection.  According to the article, there was an investigation at UC Davis into the practices of the surgeons.  It was determined by UC Davis that they did not have IRB approval to carry out the treatments and that there were some other issues with the practice going on.  At the conclusion of the investigation UC Davis wrote a letter to the FDA detailing the case and has banned the two neurosurgeons from performing medical research on humans.  Read the article for much more detail and see the link below.

Many interconnected issues in here involving IRB approval, human experimental treatments and informed consent, UC Davis, and even “probiotics”.  Still taking it all in …  Uggh …

Other stories posted in the SacBee at the same time:

See also

UPDATE 7/22/12. Some tidbits to consider
  • The doctors used the bacterium Enterobacter aerogenes for the treatment.  It wear obtained from ATCC and grown by a graduate student at UC Davis.
  • The use of the bacterium for human treatments violated the ATCC MTA.
  • It is unclear from the details here why this bacterial strain/species was selected.  But I assume it is related to the referenced Neurosurgery article (see more below).
  • The hypothesis that purposefully causing an infection may help glioblastoma patients seems to come from the observation that patients with glioblastoma w/ postoperative infections have better survival than those who do not get infections.  This could be due to many many factors jumping to purposefully causing infections with E. aerogenes seems a big big jump.
  • It would be nice to know more about the statement “Early this year, as required by University policy Drs. Muizelaar and Schrot submitted a Record of ~Invention for the bacterial intervention to UCD’s technology transfer office.” in the letter from UCD to the FDA.  Was this just a formality or were the surgeons looking to patent/protect the bacterial treatment method?
  • The Neurosurgery article discussing infection and glioblastoma may be “Long-term Remission of Malignant Brain Tumors after Intracranial Infection: A Report of Four Cases” Neurosurgery: March 1999 – Volume 44 – Issue 3 – pp 636-642.  This reported that some of the patients with infections that seemed to have a longer survival with glioblastoma were infected with E. aerogenes.
    • “In three of the cases described above, Enterobacter aerogenes was recovered from microbial cultures. Whether the presence of Enterobacter aerogenes was coincidental or whether this organism plays an important role in tumor defense is not known and cannot be proven from the cases reported. “
  • See also
UPDATE 3 – some papers on bacterial infections and glioblastoma and other cancers
  • Biocrime or a Passion to Save a Life?.  This pointed me to the article below:
  • A key article of interest: Post-operative infection may influence survival in patients with glioblastoma: simply a myth?: Glioblastoma, infection and survival from 2011. The article casts some doubt on the basis for the treatment used here
    • Citation: De Bonis P, M D AA, M D GL, de Waure C, Mangiola A, Pettorini BL, Pompucci A, Balducci M, Fiorentino A, Lauriola L, Anile C, Maira G. 2011. Post-operative infection may influence survival in patients with glioblastoma: simply a myth?: Glioblastoma, infection and survival. Neurosurgery. 2011 Oct;69(4):864-8; discussion 868-9.
    • “One of the myths that continues to be perpetrated in neurosurgery relates to the observation that a postoperative infection may actually confer a survival advantage in patients with malignant glial tumors”
    • The take-home message of this study, which can be applied to any aspect of neurosurgery, is to do everything possible to prevent a postoperative wound or cavity infection. The association between infection and prolonged survival is not definitive; we acknowledge the considerable difficulties in undertaking this type of study in a retrospective manner in view of the numerous clinical variables. A prospective randomized study on this subject is clearly not possible. Nevertheless, we believe the results of this study are important and can be used as a stimulus for further multicentric studies (to increase the number of patients) or for experimental studies using genetically modified bacteria for the treatment of GBM.
  • Also see The survival impact of postoperative infection in patients with glioblastoma multiform from 2009.
    • In this single-center study, postoperative infection did not confer any survival advantage in patients with glioblastoma multiforme”
    • Available free online here
    • “This study did not show a causal relationship between postoperative infection and prolonged survival in patients with GBM. Although targeted immunotherapy may provide antitumoral effects, simple infection does not appear to do so. Modern aseptic and antiseptic surgical techniques continue to be integral to the care of patients with gliomas.
  • Also see  Cancer J. 2012 Jan-Feb;18(1):59-68. Immunotherapy for the treatment of glioblastoma. Thomas AA, Ernstoff MS, Fadul CE. (though I cannot seem to be able to get a copy ..)

    UPDATE 6: 7/23 10 AM – some info. on UC Davis IRB, Med School, etc

      UPDATE 9: SacBee Editorial calling for Muizelear to step down as Chair
      • The Sac Bee has an editorial today calling for the doctor involved in this issue to step down as chair of the Neurosurgery department.  Some quotes below
      • Experimentation on terminal patients requires a specific set of protections for good reason. People who are, quite literally, on death’s doorstep are extremely vulnerable, and therefore not always able to give informed consent.
      • University officials conceded that “systemic issues” within the medical center may have contributed to errors made and that “additional measures designed to avoid future confusion” have been put in place.”
      • “Curiously, even after it was imposed last fall, the university named Muizelaar to fill its new Julian R. Youmans endowed chair in neurological surgery. The donor specified, university officials explained, that the chair be filled with the head of the department, a fact that begs the question: Why is Muizelaar still chairman of the department?”
      UPDATE 9: some new stories
      UPDATE 11: Muizelaar steps down as chair of department, at least temporarily
      • UC Davis neurosurgeon department chair steps down pending …
        • “One advocate for ethical human subjects research questioned the university’s decision to keep the matter in-house instead of seeking outside review. “The time is long gone for another internal investigation,” said Elizabeth Woeckner, founder and director of Citizens for Responsible Care and Research, or CIRCARE. The nonprofit group works to improve protections for human subjects in research. Woeckner called the doctors’ work on the patients – intentionally infecting them with bacteria restricted to use in lab rats only – as “the worst thing I’ve seen in my 12 years with CIRCARE.””
      UPDATE 12 – September 7, 2012 – some new news stories on Federal investigations
      UPDATE 13: December 11, 2012 – Federal investigation at UC Davis, Resignation of UC Davis Dean of Medicine, more
      UPDATE 16: August 15, 2015
      Well, many things have happened since 2012 – here are some stories to look at

      #Microbe humor: Spacewalk 1986: Bacteria

      “Who Cares About Bacteria?”

      Dubious Press Release from Cedars-Sinai linking Irritable Bowel Syndrome (IBS) and Bacteria in Gut

      Quick one here.

      Not impressed with this press release from Cedar-Sinai: Dr. Pimentel links IBS and gut bacteria – Cedars-Sinai (see other variants of it here: Daily Disruption – Cedars-Sinai Study Links Irritable Bowel Syndrome (IBS) and Bacteria in Gut and here: Irritable bowel syndrome clearly linked to gut bacteria).

      Among the things that bug me here:

      • They don’t include a link to the paper or even provide a citation
      • They claim that culturing microbes is the “gold standard” for connecting bacteria to the cause of this disease.  AND they imply this is the first method to use culturing to study the disease.  Both notions are wrongheaded.  
      • They confuse cause of IBS and symptoms.  They say that b/c antibiotics help reduce symptoms, therefore, bacteria cause the disease.  Really?  So then fevers must cause things like malaria and flu because ibuprofen helps reduce symptoms right?
      • At some point it might be nice to mention that the MD behind the new study has also been pushing the idea that IBS is caused by bacterial overgrowth for many years both in a book and via a testing company though it is unclear what his association with the company is.  I note – ads for the book claim ” In addition, Dr. Pimentel presents a simple treatment protocol that will not only help you resolve your IBS symptoms, but will also prevent their recurrence.”  So – apparently he already had a cure BEFORE the new study was even done.  I general I am skeptical of papers that show evidence for something coming from someone who apparently already “knew” the answer.

      Of course, I am not saying IBS is NOT caused by bacterial overgrowth as they claim.  But I can say this – PRs like this make me skeptical that anything new was done in this current publication.

      Diabetes & H.pylori – a correlation but no known causation despite authors claims

      Am having a hard time right now with the comments from the authors of this new paper showing a correlation between H. pylori presence and both type II diabetes and blood glucose levels.  As far as I can tell, the paper does not show any causal connection.  That is, they do not determine if H. pylori infection is a cause of blood sugar issues or a consequence of blood sugar issues.

      Yet the authors of the paper, one of whom (Martin Blaser) is a very respected H. pylori expert are saying things like

      This study provides further evidence of late-in-life cost to having H. pylori,

      And they suggest that antibiotic treatment for the elderly may help prevent diabetes.

      This to be seems to be a bit over the top.  Yes, it makes sense that H. pylori could cause these issues.  And they have a model for how it might.  But they really should be more careful with their words until a causal connection is established.  After all, we have many well known negative effects of antibiotic overuse, including some shown by Blaser.  The last thing we need is people going out and dosing up on antibiotics in the hope that it will prevent type II diabetes.  But I can guarantee that is what will happen if this story gets overplayed.

      At least a few sources report on the lack of anything showing a causal connection (e.g. see US News and World Report):

      An expert not involved with the study said that while it did not show a cause-and-effect relationship between the bacterium and diabetes, the findings suggest certain possibilities

      But I am worried that that is not enough skepticism to counteract the claims of the authors here. The study is certainly interesting.  And their model for a causal connection is fine.  But they probably need to do a little bit of toning down of their claims here.

      UPDATE: 3/17/13

      After some people asked me questions about this study at a few recent meetings I dug a little deeper.  And I am a bit startled to find out what the basis is for Chen and Blaser to claim a causative association between H. pylori and type II diabetes/ glucose levels.  Here is a summary of their logic:

      Helicobacter pylori is acquired almost exclusively in childhood [8], and there is no clear mechanism for how glucose intolerance present only after the age of 18 would increase risk of H. pylori colonization. It also is unlikely that H. pylori positivity and high levels of HbA1c levels share a mutual antecedent cause because there is no diathesis to both acquire H. pylori and to cause glucose intolerance.

      They go on to discuss other lines of indirect evidence for why they think their conclusion is correct.  And some of this is very suggestive.  But “likely” and “suggestive” is not proof.  There are many possible issues with their conclusion.  In particular I think it is easy to come up with a scenario whereby something about the host (either their genetics or their history of exposure or even their micro biome) could influence both whether or not they get colonized by H. pylori or even whether or not they get colonized by particular strains of HP.  And the same factor could influence microbiome interactions later in life.  I see no evidence to indicate that H. pylori is the causative agent here.  And for them to then basically recommend prophylactic antibiotics for elderly with HP seems dangerous at best.

      Yes, Virginia, Cell Phones Have Bacteria on Them … And this means????

      A new report is out with a discussion of microbes on cell phones: Study: Cellphones can be more germ-infested than toilet handle | News – Home. Not sure what was done in the study but regardless it seems to be focused on culture based work. And as is usually, the finding of some microbes related to ones known to cause disease leads to the inevitable conclusion that we must kill everything on the phones.
      It seems to me that we need a bit more detail on what microbes are found on cell phones before bringing out the cleaners and the irradiation and such.

      Draft blog post cleanup #1: Divide and Conquer to Find Orthologs

      OK – I am cleaning out my draft blog post list.  I start many posts and don’t finish them and then they sit in the draft section of blogger.  Well, I am going to try to clean some of that up by writing some mini posts.  Here is the first —

      Saw an interesting paper worth checking out:
      PLoS ONE: Calculating Orthologs in Bacteria and Archaea: A Divide and Conquer Approach

      It describes not only a way to speed up continual ortholog annotation in bacterial and archaeal genomes but also is linked to an ongoing open code development project.

      Here is the abstract:

      Among proteins, orthologs are defined as those that are derived by vertical descent from a single progenitor in the last common ancestor of their host organisms. Our goal is to compute a complete set of protein orthologs derived from all currently available complete bacterial and archaeal genomes. Traditional approaches typically rely on all-against-all BLAST searching which is prohibitively expensive in terms of hardware requirements or computational time (requiring an estimated 18 months or more on a typical server). Here, we present xBASE-Orth, a system for ongoing ortholog annotation, which applies a “divide and conquer” approach and adopts a pragmatic scheme that trades accuracy for speed. Starting at species level, xBASE-Orth carefully constructs and uses pan-genomes as proxies for the full collections of coding sequences at each level as it progressively climbs the taxonomic tree using the previously computed data. This leads to a significant decrease in the number of alignments that need to be performed, which translates into faster computation, making ortholog computation possible on a global scale. Using xBASE-Orth, we analyzed an NCBI collection of 1,288 bacterial and 94 archaeal complete genomes with more than 4 million coding sequences in 5 weeks and predicted more than 700 million ortholog pairs, clustered in 175,531 orthologous groups. We have also identified sets of highly conserved bacterial and archaeal orthologs and in so doing have highlighted anomalies in genome annotation and in the proposed composition of the minimal bacterial genome. In summary, our approach allows for scalable and efficient computation of the bacterial and archaeal ortholog annotations. In addition, due to its hierarchical nature, it is suitable for incorporating novel complete genomes and alternative genome annotations. The computed ortholog data and a continuously evolving set of applications based on it are integrated in the xBASE database, available at

      Definitely worth checking out.

      Twisted Tree of Life Award #12: Billion Year Old Smart Bacteria That Perfectly Treat Cancer

      OMG – for crying out loud. In the following story Billion-year-old Bacteria Could be Medical Goldmine Fox News discusses studies of marine cyanobacteria at the University of Florida. It is so wrong in so many ways I do not know where to begin. Watch the video first for layers of trouble. Then, if you dare, read the article. Among the painful parts:
      All cyanobacteria are basically lumped together into a single entity
      Cyanobacteria are the oldest organisms on earth – billions of years old – which means they must have have evolved amazing chemistry to deter predators. Wow – by the same logic – bacteria in general should be even better – because bacteria are even older than cyanobacteria. And therefore – if one focused on ALL bacteria, we should find even better predator deterring chemicals. Wait – actually – why not target all life. Surely, if cyanobacteria have perfected the art of deterring predators by the fact that they are billions of years old surely the existence of life is proof that there must be some protection against predators and therefore “living organisms” have the best deterrence systems.
      Then they make the leap from cyanobacteria surviving for billions of years by deterring predators to – wait for it – wait – hold on – be patient – wait for it – to – yes that is right – deterring “a devastating human predator – cancer.” At least they did not reveal that cancer is also billions of years old.
      And then, without any further detail, they leap from this insight to that apparently the researchers have found that the cyanobacteria make the nearly perfect anticancer drug that “has a 1-2 punch” to inhibit growth factors and receptors to be extremely potent.
      Furthermore they tell us that these cyanobacteria “are valuable because unlike similar species they are smart – targeting bad cells and sparing healthy ones.” That is right, the cyanobacteria have been smart enough to target their drugs to human cancer cells – something they must encounter frequently in their marine life.
      Oh for f3#*$# sake. I can’t even write about this anymore.
      I will just give out a well deserved “Twisted Tree of Life” award here. Not sure though who should get it – because it is unclear if this material came from U. Florida or if the station somehow came to it itself.

      Past winners include

        The story behind Pseudomonas syringae comparative genomics / pathogenicity paper; guest post by David Baltrus (@surt_lab)

        More fun from the community.  Today I am very happy to have another guest post in my “Story behind the paper” series.  This one comes to us from David Baltrus, an Assistant Professor at University of Arizona.  For more on David see his lab page here and his twitter feed here.  David has a very nice post here about a paper on the “Dynamic evolution of pathogenicity revealed by sequencing and comparative genomics of 19 Pseudomonas syringae isolates” which was published in PLoS Pathogens in July.  There is some fun/interesting stuff in the paper, including analysis of the “core” and “pan” genome of this species.  Anyway – David saw my request for posts and I am very happy that he responded.  Without further ado – here is his story (I note – I added a few links and Italics but otherwise he wrote the whole thing …).
        I first want to than Jonathan for giving me this opportunity. I am a big fan of “behind the science” stories, a habit I fed in grad school by reading every Perspectives (from the journal Genetics) article that I could get a hold of. Science can be rough, but I remember finding solace in stories about the false starts and triumphs of other researchers and how randomness and luck manage to figure into any discovery. If anything I hope to use this space to document this as it is fresh in my mind so that (inevitably) when the bad science days roll around I can have something to look back on. In the very least, I’m looking forward to mining this space in the future for quotes to prove just how little I truly understood about my research topics in 2011. It took a village to get this paper published, so apologies in advance to those that I fail to mention. Also want to mention this upfront, Marc Nishimura is my co-author and had a hand in every single aspect of this paper.
        Joining the Dangl Lab
        This project really started way back in 2006, when I interviewed for a postdoc with Jeff Dangl at UNC Chapel Hill. In grad school I had focused on understanding microbial evolution and genetics but I figured that the best use of my postdoc would be to learn and understand genomics and bioinformatics. I was just about to finish up my PhD and was lucky enough to have some choices when it came around to choosing what to do next. I actually had no clue about Dangl’s research until stumbling across one of his papers in Genetics, which gave me the impression that he was interested in bringing an evolutionary approach to studies of the plant pathogen Pseudomonas syringae. I was interested in plant pathogens because, while I wanted to study host/pathogen evolution, my grad school projects on Helicobacter pylori showed me just how much fun it is dealing with the bureaucracy of handling human pathogens. There is extensive overlap in the mechanisms of pathogenesis between plant and human pathogens, but no one really cares how many Arabidopsis plants you infect or if you dispose of them humanely (so long as the transgenes remain out of nature!). By the time I interviewed with Jeff I was leaning towards joining a different lab, but the visit to Chapel Hill went very well and by the end I was primed for Dangl’s sales pitch. This went something along the lines of “look, you can go join another lab and do excellent work that would be the same kinds of things that you did in grad school…or you can come here and be challenged by jumping into the unknown”. How can you turn that down? Jeff sold me on continuing a project started by Jeff Chang (now a PI at Oregon State), on categorizing the diversity of virulence proteins (type III effector proteins to be exact) that were translocated into hosts by the plant pathogen Pseudomonas syringae. Type III effectors are one of the main determinants of virulence in numerous gram negative plant and animal pathogens and are translocated into host cells to ultimately disrupt immune functions (I’m simplifying a lot here). Chang had already created genomic libraries and had screened through random genomic fragments of numerous P. syringae genomes to identify all of the type III effectors within 8 or so phylogenetically diverse strains. The hope was that they would find a bunch of new effectors by screening strains from different hosts. Although this method worked well for IDing potential effectors, I was under the impression that it was going to be difficult to place and verify these effectors without more genomic information. I was therefore brought in to figure out a way to sequence numerous P. syringae genomes without burning through a Scrooge McDuckian money bin worth of grant money. We had a thought that some type of grand pattern would emerge after pooling all this data but really we were taking a shot in the dark.

        Tomato leaves after 10 days infection by the tomato pathogen P.syringae DC3000 (left) as well as a less virulent strain (right). Disease symptoms are dependent on a type III secretion system.
        Moments of Randomness that Shape Science
        When I actually started the postdoc, next generation sequencing technologies were just beginning to take off. It was becoming routine to use 454 sequencing to generate bacterial genome sequences, although Sanger sequencing was still necessary to close these genomes. Dangl had it in his mind that there had to be a way to capitalize on the developing Solexa (later Illumina) technology in order to sequence P. syringae genomes. There were a couple of strokes of luck here that conspired to make this project completely worthwhile. I arrived at UNC about a year before the UNC Genome Analysis core facility came online. Sequencing runs during the early years of this core facility were subsidized by UNC, so we were able to sequence many Illumina libraries very cheaply. This gave us the opportunity to play around with sequencing options at low cost, so we could explore parameter space and find the best sequencing strategy. This also meant that I was able to learn the ins and outs of making libraries at the same time as those working in the core facility (Piotr Mieczkowski was a tremendous resource). Secondly, I started this postdoc without knowing a lick of UNIX or perl and knew that I was going to have to learn these if I had any hope of assembling and analyzing genomes. I was very lucky to have Corbin Jones and his lab 3 floors above me in the same building to help work through my kindergarden level programming skills. Corbin was really instrumental to all of these projects as well as in keeping me sane and I doubt that these projects would have turned out anywhere near as well without him. Lastly, plant pathogens in general, and P. syringae in particular, were poised to greatly benefit from next generation sequencing in 2006. While there was ample funding to completely sequence (close) genomes for numerous human pathogens, lower funding opportunities for plant pathogens meant that we were forced to be more creative if we were going to pull of sequencing a variety of P.syringae strains. This pushed us into trying a NGS approach in the first place. I suspect that it’s no coincidence that, independently of our group, the NGS assembler Velvet was first utilized for assembling P.syringae isolates.
        The Frustrations of Library Making
        Through a collaboration with Elaine Mardis’s group at Washington University St. Louis, we got some initial data back that suggested it would be difficult to make sense of bacterial genomes at that time using only Illumina (the paired end kits weren’t released until later). There simply wasn’t good enough coverage of the genome to create quality assemblies with the assemblers available at this time (SSAKE and VCAKE, our own (really Will Jeck’s) take on SSAKE). Therefore we decided to try a hybrid approach, combining low coverage 454 runs (initially separate GS Flex runs with regular reads and paired ends, and later one run with long paired ends) with Illumina reads to fill in the gaps and leveraging this data to correct for any biases inherent in the different sequencing technologies. Since there was no core facility at UNC when I started making libraries, I had to travel around in order to find the necessary equipment. The closest place that I could find a machine to precisely shear DNA was Fred Dietrich’s lab at Duke. More than a handful of mornings were spent riding a TTA bus from UNC to Duke, with a cooler full of genomic DNA on dry ice (most times having to explain to the bus drivers how I wasn’t hauling anything dangerous), spending a couple of hours on Fred’s hydroshear, then returning to UNC hoping that everything worked well. There really is no feeling like spending a half a day travelling/shearing only to find out that the genomic DNA ended up the wrong size. We were actually planning to sequence one more strain of P. syringae, and already had Illumina data, but left this one out because we filled two plates of 454 sequencing and didn’t have room for a ninth strain. In the end there were two very closely related strains (P.syringae aptata or P. syringae atrofaciens) left to make libraries for and the aptata genome sheared better on the last trip than atrofaciens. If you’ve ever wondered why researchers pick certain strains to analyze, know that sometimes it just comes down to which strain worked first. Sometimes there were problems even when the DNA was processed correctly. I initially had trouble making the 454 libraries correctly in that, although I would follow the protocol exactly, I would lose the DNA somewhere before the final step. I was able to trace down the problem to using an old (I have no clue when the Dangl lab bought it, but it looked as useable as salmon sperm ever does) bottle of salmon sperm DNA during library prep. There were also a couple of times that I successfully constructed Illumina libraries only to have the sequencing runs dominated by few actual sequences. These problems ultimately stemmed from trying to use homebrew kits (I think) for constructing Illumina libraries. Once these problems were resolved, Josie Reinhardt managed to pull everything together and create a pipeline for hybrid genome assembly and we published our first hybrid genome assembly in Genome Research. At that moment it was a thrill that we could actually assemble a genome for such a low cost. It definitely wasn’t a completely sequenced genome, but it was enough to make calls about the presence or absence of genes.
        Waiting for the story to Emerge
        There are multiple ways to perform research. We are all taught about how important it is to define testable hypothesis and to set up appropriate experiments to falsify these educated guesses. Lately, thanks to the age of genomics, it has become easier and feasible to accumulate as much genomic data as possible and find stories within that data. We took this approach with the Pseudomonas syringae genome sequences because we knew that there was going to be a wealth of information, and it was just a matter of what to focus on. Starting my postdoc I was optimistic that our sampling scheme would allow us to test questions about how host range evolves within plant pathogens (and conversely, identify the genes that control host range) because the strains we were going to sequence were all isolated from a variety of diseased hosts. My naive viewpoint was that we were going to be able to categorize virulence genes across all these strains, compare suites of virulence genes from strains that were pathogens of different hosts, and voila…we would understand host range evolution. The more I started reading about plant pathology the more I became convinced that this approach was limited. The biggest problem is that, unlike some pathogens, P. syringae can persist in a variety of environments with strains able to survive our flourish or on a variety of hosts. Sure we had strains that were known pathogens of certain host plants, but you can’t just assume that these are the only relevant hosts. Subjective definitions are not your friend when wading into the waters of genomic comparisons.
        We were quite surprised that, although type III effectors are gained and lost rapidly across P.syringae and our sequenced strains were isolated from diverse hosts, we only managed to identify a handful of new effector families. I should also mention here that Artur Romanchuk came on board and did an extensive amount of work analyzing gene repertoires across strains. A couple of nice stories did ultimately emerge by comparing gene sequences across strains and matching these up with virulence in planta (we are able to show how mutation and recombination altered two different virulence genes across strains), but my two favorite stories from this paper came about from my habit of persistently staring at genome sequences and annotations. As I said above, a major goal of this paper was to categorize the suites of a particular type of virulence gene (type III effectors) across P. syringae. I was staring at gene repertoires across strains when I noticed that two of the strains had very few of these effectors (10 or so) compared to most of the other strains (20-30). When I plotted total numbers of effectors across strains, a phylogenetic pattern arose where genomes from a subset of closely related P. syringae strains possessed lower numbers of effectors. I then got the idea to survey for other classes of virulence genes, and sure enough, strains with the lowest numbers of effectors all shared pathways for the production of well characterized toxin genes (Non ribosomal peptide synthase (NRPS) toxins are secreted out of P. syringae cells and are virulence factors, but are not translocated through the type III secretion system). One exception did arise across this handful of strains (a pea pathogen isolate from pathovar pisi) in that this strain has lost each of these conserved toxin pathways and also contain the highest number of effectors within this phylogenetic group. The relationship between effector number and toxin presence remains a correlation at the present time, but I’m excited to be able to try and figure out what this means in my own lab.
        Modified Figure 3 from the paper. Strain names are listed on the left and are color coded for phylogenetic similarity. Blue boxes indicate that the virulence gene/toxin pathway is present, green indicates that the pathway is likely present but sequence was truncated or incomplete, while box indicates absence. I have circled the group II strains, which have the lowest numbers of type III effectors while also having two conserved toxin pathways (syringomycin and syringolin). Note that the Pisi strain (Ppi R6) lacks these toxin pathways.
        The other story was a complete stroke of luck. P. syringae genomes are typically 6Mb (6 million base pairs) in size, but one strain that we sequenced (a cucumber pathogen) contained an extra 1Mb of sequence. Moreoever, the two largest assembled contigs from this strain were full of genes that weren’t present in any other P. syringae strain. After some similarity comparisons, I learned that there was a small bit of overlap between each of these contigs and performed PCR to confirm this. Then, as a hunch, I designed primers facing out of each end of the contig and was able to confirm that this extra 1Mb of sequence was circular in conformation and likely separate from the chromosome. I got a bit lucky here because there was a small bit (500bp or so) of sequence that was not assembled with either of these two contigs that closed the circle (a lot more and I wouldn’t have gotten the PCR to work at all). We quickly obtained 3 other closely related strains and were able to show that only a subset of strains contain this extra 1Mb and that it doesn’t appear to be directly involved in virulence on cucumber. So it turns out that a small number (2 so far) of P. syringae strains have acquired and extra 1Mb of DNA, and we don’t quite know what any of these ~700 extra genes do. There are no obvious pathways present aside from additional chromosomal maintenance genes, extra tRNAs in the same ratio as the chromosomal copies, and a couple of secretion systems. So somehow we managed to randomly pick the right strain to capture a very recent event that increased the genome size of this one strain by 15% or so. We’ve made some headway on this megaplasmid story since I started my lab, but I’ll save that for future blog posts.
        Modified Figure S12 from the paper. Strains that contain the 1Mb megaplasmid (Pla7512 and Pla107) are slightly less virulent during growth in cucumber than strains lacking the megaplasmid (PlaYM8003, PlaYM7902). This growth defect is also measurable in vitro. In case you are wondering, I used blue and yellow because those were the dolors of my undergrad university, the University of Delaware.Reviewer Critiques
        We finally managed to get this manuscript written up by the summer of 2010 and submitted it to PLoS Biology. I figured that (as always) it would take a bit of work to address reviewer’s critiques, but we would nonetheless be able to publish without great difficulty. I was at a conference on P. syringae at Oxford in August of 2010 when I got the reviews back and learned that our paper had gotten rejected. Everyone has stories about reviewer comments and so I’d like to share one of my own favorites thus far. I don’t think it ever gets easier to read reviews when your paper has been rejected, but I was knocked back the main critique of one reviewer:

        “I realize that the investigators might not typically work in the field of bacterial genomics, but when looking at divergent strains (as opposed to resequencing to uncover SNPs among strains) it is really necessary to have complete, not draft, genomes. I realize that this might sound like a lot to ask, but if they look at comparisons of, for example, bacterial core and pan-genomes, such as the other paper on this that they cite (and numerous other examples exist), they are based on complete genome sequences. If this group does not wish to come up to the standards applied to even the most conventional bacterial genomics paper, it is their prerogative; however, they should be aware of the expectations of researchers in this field.”

        So this reviewer was basically asking us to spend an extra 50k to finish the genomes for these strains before they were scientifically useful. Although I do understand the point, this paper was never about getting things perfect but about demonstrating what is possible with draft genomes. I took the part about working in the field of bacterial genomics a bit personally I have to admit, c’mon that’s harsh, but I got over that feeling by downing a few pints in Oxford with other researchers that (judging by their research and interest in NGS) also failed to grasp the importance of spending time and money to close P. syringae genomes. We managed to rewrite this paper to address most of the other reviewers critiques and finally were able to submit to PLoS Pathogens.

        Crosspost from A very misleading “bacteria in buildings” advertisement presented as “news”

        Am crossposting this from where I posted it earlier. See original post here: A very misleading “bacteria in buildings” advertisement presented as “news”
        Wow this “story” (which is really an ad) is just so incredibly bad I do not know what to say: Dangerous Bacteria Isolated in Healthcare HVAC Evaporator Coils. I do not even know where to begin with criticism so I will just go step by step through some of the advertisement.
        1. Title: Dangerous Bacteria Isolated in Healthcare HVAC Evaporator Coils
        There is no evidence that the bacteria being looked at here are dangerous.
        2. First sentence ”A recent study suggests that doctors may want to monitor the environmental condition of their air conditioners evaporator coil before surgery to help prevent the spread of bacterial infections”
        No evidence is presented anywhere that monitoring AC coils has any even remote potential value here.
        3. Second sentence: Dr. Rajiv Sahay, Laboratory Director at Environmental Diagnostics Laboratory (EDLab) and his colleagues sampled evaporator coils in healthcare air handling systems and isolated Pseudomonas aeruginosa a known noscocomial pathogen.
        Well, Pseudomonas aeruginosa is indeed a known pathogen.  However, there is no evidence presented that all the things they detect are indeed pathogenic/virulent.  In fact, later in the article they report their results as being for “Pseudomonas sp” which suggests that their typing was very broad.  It is very possible that many of the cells they detected are not pathogenic.
        4. Ignore the middle part.  It is just saying that Pseudomonas aeruginosa can be nasty in compromised patients.
        5. They then go on to discuss their study more “In the study, over 560,000 colony forming units (CFU)/gram of Pseudomonas sp were isolated from deep within the evaporator coil system.”
        What study?  No data is presented.  No methods.  No results.  Nothing.
        6. They then say “Potential aerosolization of these micro-organisms from the infested coil is immense due to a discharge of air stream with 6 miles/hours (commonly observed) across the evaporator coils”
        Not so sure about that.  Would have been much better to study ACTUAL aerosolization.
        7. Then we find out that they person who conducted the study Dr. Rajiv Sahay is also the one selling the cleaning service to clean your air coils.  That does not instill confidence in me.
        So a person selling HVAC cleaning reports unpublished results that they claim suggest if you do not clean your HVACs in hospitals you put all your patients at risk.  I am on board with the need to study microbes in hospitals more.  I am on board with the potential risks of microbes in AC systems.  I am not on board with not presenting data, and with getting the science wrong.