Tip:
Highlight text to annotate it
X
Dan Roden: This is a refined circle, okay? And I think
that what it sort of emphasizes is this idea that there's, you know, so no one else in
this room is an arrhythmia guy, so you don't know what figure of eight reentry is, but
it's a big -- it's a big deal in my world, and this is the one reentrant loop that you
actually don't want to interrupt, but -- so I leave it at that.
There's this feedback that goes on all the time between clinical and basic research,
and between, and that goes on in the electronic record environment as well as anywhere else,
so plans for my meeting [spelled phonetically] was the real slide.
Teri [spelled phonetically] and I and a bunch of other people thought about this problem
last night, and we thought about themes for the meeting. We thought about, so, what are
the barriers to executing this vision of genomic medicine that we all sort of seem to have
in, at the 50,000 and 500,000 foot view, we sort of share it, and when you get down to
the 1,000 foot view, we all have different ideas of what that vision actually looks like,
but the notion was to focus the next meeting on barriers to executing that kind of vision.
So, Ledbetter [spelled phonetically] navigating -- it's this one. So, here's the slide that
Teri thankfully made up for me. The original theme was going to be standardizing formats,
and we talked it through last night. We've talked it through today, and I don't think
we're ready to have a day and a half long meeting on standardizing formats, and these
are -- so this is a proposal, this is not, I don't think this is written in stone. This
has been bothering me this entire meeting. The electricity keeps on bouncing back and
forth on this laptop.
So, the notion was to focus on the barriers and how we might address one of the barriers
that we keep on hearing about, but I don't think we have a good sense of how that, how
it really works in real time as the payers. I added the pharmacy benefits people to this
list after we had spoken last night, and then perhaps CMS is sort of the ultimate payer.
We talk a lot about CLIA and what that means, but we thought it might be useful to hear
from people who actually set those standards, and then some of the genomic reference expertise
as a way of dealing with, again, another barrier to incorporating genomic information into
the flow of clinical healthcare. The notion of seeing some early deliverables at the May
meeting is written down on this slide. I won't say who wrote it down because we all take,
we all -- I'm standing it here presenting it, so I have to take some credit for it.
I already see Debbie [spelled phonetically] nodding her head no.
Female Speaker: No, no, I just [unintelligible]
Dan Roden: [laughs] Well, so, I think the way to think
about this is to think about what it is we want to accomplish as a group, and what it
is that -- what we could do is to think about the barriers to executing the visions that
we have and then deal with them. So, that was the proposal, and I think I'm open to
comment. Boy. For a -- I was about to say, not even Mark Ratain [spelled phonetically]
has anything to say, but I spoke too soon. Mark.
Mark Ratain: I would be interested in hearing from experts
in genomics law.
Dan Roden: So, maybe from the University of Wisconsin
to start with, right?
Mark Ratain: I don't know.
Dan Roden: [laughs] Okay. So, I'm going to actually --
Mark Ratain: So, what do you mean by that? There's actually
apparently a field --
Female Speaker: There is a field.
Male Speaker: And what [unintelligible] would you like to
hear about?
Mark Ratain: Well, I would like to hear, for example, if
there are research, say there's a research whole genome sequence, okay, and a patient
in a study, but their physician potentially has access to that information, is that information,
is that data, if that individual tried to apply for disability insurance, is that genomic
information, will a staff need to turn over to the insurance company?
Male Speaker: Yeah. Dan, I mean, I'm getting back to the
awesome Tiffany [spelled phonetically] standards, and the question really is, we had come up
with an awful lot, you know, in terms of not only the sequencing data analysis but also
and what is actually quite a substantial undertaking, as you know, the phenotype standards. Did
you include the phenotype? Is that included in the Tiffany standards, and how should we
think about this between now and May?
Male Speaker: And I'll answer that question.
Dan Roden: Ah, okay. Good, because I'm glad somebody's
going to answer.
Male Speaker: In a minute.
Dan Roden: Oh, so, and the, well, you were in the room
last night. So, we talked a lot about that, and it wasn't a major focus of Howard's [spelled
phonetically] talk this morning, but I think it is part of what we want to get accomplished.
Now that that said, I'm not sure that Tiffany standards, and I'm not sure where that, I'm
not even sure where that word came from, although we talked about it --
[talking simultaneously]
Yeah, I know, although the --
Male Speaker: It should be the dungeon [spelled phonetically]
standards.
Dan Roden: I get it. I get it. I get it. I get it. For
those of you who weren't in the room, it was not a well-lit bright room with a table. It
was a little dungeon down in the third subbasement with a bunch of chairs around the edges, and
bad fluorescent lighting, and that was it. No pictures on the walls or anything.
Female Speaker: [unintelligible] much better --
[laughter]
Dan Roden: Well, next meeting, next time we're going
to meet in the ladies room. So, the Tiffany standards. Okay. Now I get it. You know, would
probably include a section on phenotyping. But I guess if we're going to develop those
kinds of standards, somebody needs to start to work on them, and then they can present
to us wherever they are when they are there that --
Male Speaker: And I will talk about that next step, so we
Dan Roden: Debbie [spelled phonetically]?
Female Speaker: I moved so it wouldn't be so [spelled phonetically]
hard. Yes. I think there are a couple of things here. I mean, I guess CLIA and CAP certification
and experts, it would be interesting to engage them in the Tiffany standards kind of thing,
I mean, because you develop a set of standards, if they don't meet any guidelines, then you're
in trouble, and we've heard that some people are considering the laboratory guidelines,
but maybe not the analysis guidelines. So, I think it would be good to have some of this
discussion ahead of time.
The other thing is I'm not really sure I understand engage thousands and [spelled phonetically]
genome reference. Is this genome centers? I just want to be more precise on that.
[talking simultaneously]
Yeah.
Male Speaker: Nist [spelled phonetically].
Male Speaker: Yeah. So, I thought the idea came up in the
breakout last night, and also I was talking to Howard before. Just, it's not, it's -- some
of it's genome center expertise, but some of it's analysts that are with various consortias.
So, for example, a thousand genomes analysis group have, they have variant collars [spelled
phonetically] that they've actually compared to each other, and they actually understand
the performance. And Howard said, "Well, [unintelligible] --"
Female Speaker: So --
[laughter]
So, I would make it more specific.
Male Speaker: You had your chance.
Female Speaker: I was like really weird. I would make it more
specific. So, experts engage experts on base-calling indel and CMV calling, and what is the reference
sequence that all -- that everything's compared to?
Male Speaker: Should I call it variant calling?
Female Speaker: That's more, it's just more specific. That's
all, so you know who you're going after.
Male Speaker: If I make a spelling mistake, that's because
I'm [unintelligible].
Female Speaker: So, Kelly [spelled phonetically].
Male Speaker: This is fun.
[talking simultaneously]
Female Speaker: I might even make this a little more specific.
So, I think like these 10 genomes that people are talking about, I mean, I think one of
the things that we've discussed at length and various sections of this meeting is the
fact that we want to be able to combine data sets from numerous places that are generating,
you know, cancer genomes for diagnostic purposes, for treatment purposes, and if we could get
a set of 10 genomes that we would ask everybody to include, so if they're using the Illumina
platform, it would be 10 Illumina platform genomes. If they're using the solid [spelled
phonetically], then it would be 10 solid platform genomes that they would include in their analysis,
so that eventually there will be the collection of these from various sites, I imagine, at
some point in the future. And then when we -- it's going to be difficult to take that
information that we get from people and standardize it, but if everybody had run through the same
10 genomes through their pipelines, then we would have at least an idea of where missing
data and stuff of that nature might be in their analysis.
Male Speaker: [unintelligible] more than just genomes, are
we going to do rough C [spelled phonetically]?
Dan Roden: Yeah, so, I mean, this is really a recommendation
of the sequencing working group, and if I could get my computer to work, I'm having
problems, but let me just try to see if I can get my thing loaded up here, and then
address that question.
Male Speaker: So, I'm just going to save this, and --
Dan Roden:? Save that, yeah.
Male Speaker: -- and go away.
Dan Roden:? We shouldn't have gone out of order. That
messed things up.
Male Speaker: That's okay.
Female Speaker: [unintelligible]
Male Speaker: Yeah, just in this [spelled phonetically]
CLIA and CAP certification experts. I'm happy to have them attend, but the AMP [spelled
phonetically] and ACMG laboratory guideline development groups around exome sequencing,
whole genome sequencing, are probably the more expert group, so you want to include
them and often, there's overlap in the individual people. AMP?
Male Speaker: Wow.
Male Speaker: Sorry [spelled phonetically].
Female Speaker: So, then, all --
Male Speaker: You got all this down [spelled phonetically].
Okay.
Male Speaker: Yeah, and ACMG, in addition to CLIA/CAP.
[talking simultaneously]
Female Speaker: [unintelligible] better staffing [unintelligible]
Male Speaker: Okay.
Female Speaker: David, you want to, no, that was David.
Male Speaker: So, towards the issue about the 1000 Genome
Project and to follow what Adam [spelled phonetically] was talking about, this comes out of a conversation
Scott and I had last night in the bar after you all went to bed, so it was alcohol-induced,
but the issue was one of the things that's happing as part of the 1000 Genome Project
and maybe other discussions, is that there's weekly discussions every week about analysis
strategies, and so for those of us that are interested in leveraging that information,
we'd like to be sitting right behind that.
So, as those analysis strategies get firmed up, then I'd love to be in a position where
we could use those in a clinical type laboratory, and so I think that was one of the discussions
around what Adam was saying is that how do we roll this down the pipeline so that it
goes from 1000 Genome Project type. It doesn't have to be just 1,000 genomes, down towards
clinical.
Female Speaker: David first, please.
Male Speaker: So, on the 1000 Genome side, there's weekly
calls do have comparisons, and what happens is immediately people who don't do as well
start to make changes. And so one of the things that you emphasized was [unintelligible] 1.8
is 1.8, whereas GA2K [spelled phonetically] has, you know, you could do a nightly amp
bell [spelled phonetically] and the versions change quite a bit. And them getting the message
of versioning and some of the needs that you might have would be something that's important,
because I don't think they think about what you might be wanting on a clinical side, and
they think nightly builds [spelled phonetically] are great.
Male Speaker: So that's, I think that's one of the issues
is that so that if we're not involved in the discussion, there's not some freeze points
that could be then evolved downstream that could be used. So, continual evolution is
great, but if there could be some freezes, just like anything else that we freeze in
the genome project, that then enables those of us that want to start implementing having
that. So, some discussion around how to do that, I think, is what we were trying to capture.
Male Speaker: I think there are certain things like SNP
calling that's already very good on, say, GATK and they could freeze that. Indel development,
you know, is maybe 80 percent accurate right now, and we're going to have to keep changing
it until it's 100 percent. And then as fees [spelled phonetically] is even, you know,
further away from that.
Female Speaker: I think also the kinds of things you're interested
in are not, the genomics community is calling on a thousand individuals at once and getting
probabilities of getting genotypes at low pass. That's different than taking the entire
genome of an individual and calling it in the way that you're suggesting for all variants.
So, I think we need to be pretty clear on what we're asking for information on, but
it's not, doesn't mean that those things can't inform this calling. But, I do think that
it is a different issue.
Male Speaker: I totally agree.