• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Animal & Veterinary

  • Print
  • Share
  • E-mail

Session VI Questions and Answers Session

DR. CARATTOLI: Alessandra Carattoli, Rome Italy. I have a question for Maria Karlsson. I was interested in one of your data. You showed that among the top serotypes producing plasmid-mediated quinolone resistance there was the serotype Corvalis. Because one of the few examples in Europe demonstrated of an importation of meat, was associated with this Corvalis producing qnr. It has been demonstrated first in Denmark and then confirmed in the --- so have you have just the time to check if this Corvalis producing qnr --- also arrived here? If this strain was a clonal strain with those described in Europe, I mean this Corvalis.

DR. KARLSSON: No, I am sorry. I don’t have the answer to that question and that would certainly be something that we could look into and see if they are clonal and if they have the same pattern as the European isolate. Thanks.

DR. KATHARIOU: Sophia Kathariou from North Carolina State. I have a question for Matthew about the Campylobacter subtyping system. So one of the really big advantages of MLST or any sequence-based system is the portability of the data because the data means exactly the same thing no matter who obtains them. My question is, with PCR data even though you mentioned that those were SNP free, you know there is always the possibility that the new isolate might have a SNP that might make that PCR product difficult to obtain. I am wondering if the portability issue has been looked at? Thanks.

DR. GILMOUR: I actually agree, certainly with the MLST one of the strongest benefits is that you are actually getting sequence data back to analyze. 

You have “A’s”, “C’s”, “G’s”, and “T’s” to put through all your other tools that you want. Even as far as the possibility of nonspecific PCR, it is probably the caveat of CGF. I agree too, that it is just a band on a gel. There is nothing that says it is as intended. You know, if you are looking for an absence of a band, there is a strong possibility of false negatives. The presence of a band could also be a false positive. 

So I agree, but it is a trade off in throughput I guess. Because, again, the goal here is you can imagine other genomics projects, whether they are Campylobacter or Listeria or Shiga toxin-producing E. coli that you have sequenced the genome. You found 100 traits that you love and now you want to go into your freezers and screen the rest of your samples for those same traits and you need a relatively high-throughput method to do it. These multiplex assays which are actually pretty cheap to run -- they don’t even need specialized tac (sic), it is just a generic tac. So it is a good option but with those caveats that it is just a band on a gel.

(Inaudible voice)

DR. GILMOUR: So the question was did C. coli cluster separately from jejuni? Again, that is where I wish they had some figures rather than just 32 tables. But I did ask them before I left and I think there was a separation. You could tell a coli from a jejuni in general based that.

DR. ZHAO: I have a question for Matthew. Just for my curiosity, you calculated the symptom diversity to compare the other methods of, you know, discriminatory power. Have you used a tween* line for PFGE to calculate to compare symptom diversity?

DR. GILMOUR: I am pretty sure they just had the sma-1 data. And even for that it wasn’t on the full 493 isolates. It was on about 150 isolates for the PFGE.

DR. ZHAO: ---

DR. GILMOUR: No. No two enzymes.

DR. FITZGERALD: One more question for you Matt. So how many genome sequences were originally compared to come up with the loci in the beginning? Do you know?

DR. GILMOUR: It was actually created principally out of comparative genomic hybridization data from the microarrays. It would have been just on the handful of jejuni  that would have been available at the time. 

I know one of their current efforts is actually to, you know, sample amongst CGF types and MLST types through genomics, because obviously the microarrays are limited to that one or two complete genomes that were out there; you know, the content that you are screening for rather than the content of the actual isolate.

DR. FITZGERALD:  I just want to make a few more comments. Certainly from the public health side, we are looking for typing methods that give us diversity but we also need methods that give us epidemiologically relevant information. So I think it is very important to characterize, you know, to look at well-characterized strains and it certainly is an interesting method and we can certainly work with you to see how it works.

Just one other comment on the MLST data, I mean, I think that comparing our data to Europe’s and to New Zealand’s, we have done some preliminary comparison of our MLST data to what you all see in Canada. 

I just want to comment that we can’t assume that what they are finding in Europe and New Zealand is what we are going to see here in the U.S. and in Canada. Our data really is quite similar to what you all see in Canada, so I don’t think we can blame it all on poultry. I think the contributions of the other sources will be quite important here in the U.S. So it is not all poultry.

DR. KWAN: Patrick Kwan from CDC. I would just like to make a note about the MLST. When they first developed it in Oxford, I think they looked at more than just the seven housekeeping genes. I think these are the ones that they have selected out and they have -- basically when they looked at like I think about 20 loci, I mean the isolates also sort of grouped together in a similar way as if it was just by using seven. 

From the CJH data you have got, I just wonder if it will inevitably include some hypervariable regions so that might actually disrupt the general picture when you, you know -- especially in terms of talking about source attribution.

DR. FRYE: I think we are going to have to cut it off there or we will not have any lunch. One more? One more. Okay. Jean?

DR. WHICHARD: Quickly? You know not what you ask. Jean Whichard, CDC. I am very thankful for this presentation about CGF and I like to see that new methods are being developed and evaluated, and particularly that there are now objective criteria for understanding the diversity that they give and the Wallace Coefficient. I think that I probably need, you know, Statistics on Subtyping Methods for Dummy’s book in order to understand all these things. But I am just wondering, what forum do you all think is appropriate for deciding, you know, what methods or even what combination of methods?  Shaouha Zhao has looked at some combinations of various subtyping methods for the various questions being asked. 

I mean we have certain questions in the context of an outbreak and then we have other questions in terms of population genetics. You know, where are those decisions ultimately going to be made? Anyone? CHRO*, next time around?

DR. FITZGERALD: I will just say one quick thing. I mean, I think ultimately it all depends on the questions that you are asking. I mean you have got to put it into perspective and you can’t type for typing’s sake. We need to come up with some formed questions and then we can decide on which is the way to approach it, in a nutshell so we can eat lunch.

DR. McDERMOTT: I just wanted to say I know we are running late, but let’s just take an hour for lunch and come back at 12:20.

(Whereupon, a luncheon recess was taken.)