 So I think now we're going to go to the wrap up. Before we do, though, I just wanted to respond to this last discussion here about trust. And I wouldn't want to downplay that importance of that factor. But I think it probably also matters that at that community clinic, there were people getting health care, right? And so in a lot of underserved communities, whether they are poor and white, whether they are Native American, no matter what, people just aren't getting health care. They aren't getting it consistently. And so long as we are recruiting through a health care system, I mean, I think that matters as much as anything, right? All right, so just to wrap up here, I think starting out with Jim, we sort of heard that if you build it, they don't necessarily come. And you actually have to work at it. And that there are social justice reasons to work at it and scientific reasons that I think Carlos went into some detail regarding the scientific reasons. And I just wanted to add something with respect to social justice, sort of connecting this session to the previous session, right? That there are all kinds of activities that patients, at least some of them, and participants are beginning to engage in using genomic information. And we can be creating a kind of disparity right there in terms of people who can go on the internet, people who can see, oh, somebody else has a variant like mine if we're only including some populations in these studies, right? So right there is a kind of disparity that we could be creating. And that is a social justice issue. I think Carlos gave us some really great scientific arguments. But I think he pointed also to the broader set of challenges which is there are good scientific reasons for including diverse populations. How do we actually design the studies to do that most effectively so that it will both achieve social justice ends, but also achieve our scientific ends in the most economic way possible. How do we design those studies? How can they be properly powered? Can we modify some existing protocols to bring in or oversample more diverse populations? And he sort of as a side note pointed out something that of course to me is very important which is also what additionally will we learn in the LC world when we start bringing more diverse populations into these studies. Let's see what else I wanted to say. I think we heard very importantly that some of the socioeconomic barriers, again, I think these not all communities of color are not all racially and ethnically diverse communities experience the socioeconomic barriers. And some non-minority, some non-minority communities also experience these kinds of economic barriers and they're really important. We need to pay attention to them. The point that people actually lose money if they're being paid hourly that they could lose their job by taking time off to come to us. I mean, this has to do with how we design our studies when we can be available to see people and so forth. I think that seems to be crucially important point that came out. And finally, we had heard some emphasis on increasing the diversity of the workforce. And the point that you can, somebody doesn't have to be from the same community that you're aiming to study in order to have a trusting and productive relationship. On the other hand, it probably, it can help to generate a trusting and productive relationship. It can also help to improve the kinds of scientific questions or not improve them but to help us ask scientific questions that are of interest to a broader group of people in our society. The last thing I would say is that I think we didn't quite do justice or entirely do justice to our topic because we were given a very wide topic for this panel which included clinical diversity. I think broadening the group of participants, broadening to more underrepresented communities and so forth may also broaden the clinical representation but there are many ways of thinking of diversity and we only touched on a couple, thanks. So next we have our final panel on healthcare utilization, economics and value co-moderated by Katrina Armstrong and Sharon Plon. We've talked by Dave Vinstra and Pat Deverk. Maybe while we're getting slides up there's one other comment that I was thinking of and it didn't come up in our discussion but it has to do with the fact that as we move towards more studies that use electronic health records, if underrepresented communities are served by healthcare systems that don't have as good of electronic record systems that don't have as much data in their electronic record system, that also might result in people from underserved communities being systematically excluded because the data quality just might not be as good. So I think that's another thing that we really might think about going forward is how do we help to ensure that just data quality considerations don't systematically drop out people from lower income communities or underrepresented communities.