 Welcome back, everyone. I'm Elsie Sunderland, a professor at Harvard University and a member of the planning committee. In this session, we'll discuss cross-cutting issues and PFAS risk assessment. The session will examine the challenges and strategies and considering mixtures and developing class-based approaches for risk assessment. For these topics, we will have a short overview presentation intended to describe the current state of the science, what is likely to be known in the near future based on ongoing research and remaining research data gaps, followed by discussion with other federal agency experts and planning committee members. We will also discuss research needs to address emerging issues, including approaches and tools to rapidly respond to newly identified and next generation PFAS, such as new strategies for chemical screening and new in vitro and computational toxicology approaches. The first discussion will address issues associated with PFAS mixtures with an overview presentation provided by Moiz Mamta, science advisor in the Office of the Associate Director of the Agency for Toxic Substances and Disease Registry. Please take it away, Moiz. Good afternoon. I want to thank the National Academy of Science for the opportunity to talk today. PFAS exposures are unintended, complex, simultaneous, or sequential multi-route. So assessment of something like that, you would expect, is very important but also very complex. The community across the United States are concerned about potential adverse effects of PFAS exposures. We evaluate the toxicity of such exposures to advise the community or the workers that no harm is expected of such exposures. However, sometimes it is to bring awareness in the community of potential effects if established limits are exceeded. And that comes around at public meetings, health education programs for physicians, health workers, and through fact sheets. And so for this purpose, federal agencies develop methods, guidelines, framework, criteria, standards, or limits to protect human and environmental health. So single chemical assessment may underestimate the total risk. And for that reason, we need to have mixtures, methods. And those are reliable and can be easily conducted. One such example, and so the federal agencies develop all kinds of this criteria. And ATSDR 2018 Mixtures Framework is such an example. And it provides three different approaches, the whole mixture or similar mixture or component-based approach, which is called the Hazard Index approach. But you all have heard yesterday and today that how the database is not as robust as one would like to for the mixtures risk assessment. But the HIE approach, the Hazard Index approach uses the MRLs, RFDs to derive value based on some toxicity of components. There's a lot of uncertainty involved. And to resolve that uncertainty, if we go to the next slide, we'll see that yesterday, Rusty, Thomas, Mike DeVito, Michelle, several other people presented that all kinds of research are being conducted that will yield results. And so these studies are both in vitro and in vivo. And you heard this morning that the idea of limited in vivo studies to establish the concentrations, to establish the dose, to establish what is the toxic variety we conduct limited in vivo studies. But the field of psychology is moving more towards in vitro and in serico work. And so a lot of data will be coming out to fill some of this data gas, particularly the understanding of absorption, distribution, metabolism, and excretion of these compounds. So it is practically impossible to experimentally test all the possible combination of PFASs and mixtures containing PFASs. So we have to look at computational in silico tools as Rusty talked yesterday. And if you go to the next slide, then we see there's a lot of information that is needed in terms of composition of the PFAS mixtures, of the co-occurring chemicals. If we need some kind of database to establish a registry to see what kind of exposures are occurring in the environment. So that we can look at in silico methods how we can develop them. And this morning through studies which have been presented, you saw that we need biomarkers of toxicity, of exposure, of health effects. We need to understand the mechanisms of action at the cellular level, at the organ level, at the system level to understand how to evaluate combined toxicity of chemicals we are exposed to. And of course, we'll also discuss a little bit about variations in the populations, whether they are adult, whether they're children, their susceptible populations. And all that plays a role in the risk assessment and the toxicity assessment of mixtures. So all you have heard in the last couple of days, all adds up here and mixtures risk assessment is where everything is integrated and we'll talk more about it. Thank you so much for your attention. Great, thank you so much. So now we will bring in our additional federal agency discussants and Herman Gibbs, Scott Bartel, and Tom Webster from the panel committee for follow-up discussion and questions. And maybe I will hand it over to Herman to ask the first question. Okay, thank you, Elsie, I appreciate that. So how well do we understand PFAS mixtures associated with different exposure sources? I'm sorry, Moise, go ahead. Moise, we've been together for 30 years and we're graying as the time goes by. Go ahead, Moise. Jim, Jim, go ahead, please. I was going to say it sounds like a question directed to the Department of Defense. And I say that because we're probably the largest user of A-Triple-F, which has historically contained PFAS compounds. And so we're, as a group, we're very interested in finding a substitute for A-Triple-F that does not contain PFAS compounds. And we're keenly aware of the problems that our historical use of A-Triple-F has caused. And so we're keenly interested in the research being done at EPA and NTP regarding some of the newer replacement A-Triple-F mixtures coming out. Moise gave you three different methods for going after these mixtures. And unfortunately, we just don't have the basis for looking at two of those. We have to test the entire mixture in an in vitro or an in vivo test system to get at where the toxicity is. We'd certainly like to be able to do differently, but we're trying to come up with a substitute A-Triple-F formulation now, not three, four, five, 10 years from now. Go ahead, Moise. So Herman, I think it's a very interesting question. And we have heard that exposure to PFASs are from multiple sources, from drinking water, diet, dust, and other pathways. So when we are evaluating mixtures, it is important that we consider how much of it is coming through a particular source. And it goes back to the exposure assessment and how it is done. And basically the better characterization of exposure, the better problem definition for mixtures assessment and toxicity assessment. And as Jim pointed out, that is really critical. And if you look at drinking water, every state gives a different, what is called a relative contribution to a source. And when EPA and ATSDR does defined limits, it is considering the whole integrated exposure for that chemical or chemical. So we have to keep that in mind. The multiple sources are all affected by our lifestyle. Fish advisories are developed for a purpose and the levels are defined as how much in a filet, for example, in certain cultures, they use the skin. And there's a lot of persistent organic pollutants, including PFASs in there. So when we go and try to explain them, they have to change their habits of cooking and eating and all that. So it's something we have to keep in mind. It's a very complex issue. Yeah, and so when we think about, say, AFFF as one source and these complex other sources that we have that you've just mentioned, Moise, so things like seafood and drinking water, I think the question is really, can we look at the abundance of both the legacy PFAS and the new PFAS in those exposure vectors and use that in some of the in vitro testing to better understand mixtures? And to what extent is the federal government thinking about linking exposure assessment and effects in this way? So for example, from some of our research, we can see that seafood is associated with a very specific signature of long chain PFAS because those are the more bioaccumulative compounds. So you see them both as a signature in the exposure source and you can also see that signature of seafood consumption in people. So when you try to, this is a cross-cutting session. So when we try to bring it all the way to health effects and we're thinking about mixtures, can we use that exposure information to better understand outcomes associated or characteristic of specific exposure sources or what research would we need to be able to do that? That's the question really and it's really gonna depend on exposure. So if you're living next to a navy base that has contaminated groundwater, clearly that's your primary exposure and it dwarfs everything else. And so we feel the need to get in there, remediate that as quickly as possible. But on the other hand, if you're a normal American and you're being exposed to PFAS from all different directions but not necessarily a release site, those issues relating to where you get your PFAS from in your diet become critically important because they can be a significant part of your exposure. And I've seen a couple of papers here in the last six months where people are trying to tease apart that issue, at least for normal Americans where they're looking at the contribution of PFAS, total PFAS to the exposure. And they've not begun to look at individual PFAS that might come from a particular source as being critical either for health outcomes or for its contribution to the total PFAS exposure. So great questions, huge unknown as far as I'm concerned. Okay, well, something for us to all think about. I'm gonna hand the next question to Scott, if you would. Okay, so for other chemical classes such as dioxins the use of a toxic equivalency factor has proven useful for risk assessment. And you may have noticed that some health advisories and actually a handful of research papers and studies as well have already started using a sum of PFAS concentrations without trying to weight according to relative toxicity. At the same time, we have one or two interesting papers such as one by Melissa Gomez came out a couple of years ago trying to rank relative potency of different PFAS and PFAS replacements by liver toxicity and pharmacokinetic characteristics. And so these kind of seem to be moving into direction of maybe more of like a TEFT Q approach. And so I'm curious what the federal panelists think about that and what barriers there might be towards implementing that kind of approach for PFAS chemicals. I hate to be talking all the time. I feel like Susan Fenton, if I could address that. A good example would be the combination of dioxins and PCBs. So PCBs generally are looked at as dioxin-like or non-doxin-like. And so you get into trouble when you separate out PCBs in that fashion because you separate out some of the PCBs that are actually causing immunologic effects, causing reproductive effects, causing ecological effects. They do not have dioxin-like qualities. And so if you have a group of compounds that all act on a particular receptor, and for dioxin it's the AA receptor, it's presumably the key step in having dioxin toxicity. Although I would argue that it's not the case. Then you're good. But I think for PFAS would be thousands of compounds we have potentially in front of us. We have yet to define a mechanism of action that is that simplistic. It may be that PFAS have two different actions. One's receptor-based, and I would say multiple receptor-based. And the other might be membrane, fluidity-based. In other words, they act like fatty acids and become part of the membrane. I think that's gonna be a very complicated and tough topic to get our hands around. And I don't think we have enough information to say that the various PFAS compounds have a similar mechanism of action. We certainly know that's not the case for PFOA and PFAS. So even though they are both developmental toxicants, they do not have the same effect in the same species. I would agree with Jim that it seems like there's plenty of evidence in the literature that there are at least multiple nuclear receptors that are activated across various PFAS, whether carboxylates or sulfonates, since they're the two most studied subclasses. But kind of a difference with, you know, that historically TEQ or TEF, like Jim was saying, generally applies to aerohydrocarbon receptor activation by dioxins and dioxin-like compounds producing a common end point. And I don't know if that's gonna be possible with PFAS given what's already known about multiple molecular initiating events and how those can be differentially effective across different life stages, whether it's gestational exposures or adult or juvenile. And so trying to apply that type of approach really depends on kind of the backend or the apical effect or the adverse effect and being able to scale the different compounds based on the relative potencies, which is, you know, it's another similar approach to TEQ or TEF, but it's not implicit that they're all working via the same molecular mechanism or even necessarily all the same key events. And like you said, with PFAS and PFOA, you do see developmental effects, but the key events that get you to that point are quite different. Would anyone else like to? Can you hear me? Yes, please. Yeah, this is Hanai. I'm sorry, I have problems with the internet. I'm just on the phone. I just wanted to add, you know, there are two interesting studies about relative potency factors. One was from the Netherlands and the other one is Shinikario from ATSBR. And the one from the Netherlands was studying like 16 PFAS, but like you mentioned before, it's very limited. It's limited to red, male or intermediate exposure. And they used PFAS as an index chemical and they looked at benchmark doses and how these benchmark doses are related with relative potency to PFAS. But when they actually looked at the hepatic effects, it was literally absolute and relative and hepatomegaly, they were thinking about PPAR alpha. And that's really limited to humans. Although they wanted to disregard this comment in the paper, I don't really know how relevant it is to human risk assessment. The other paper from Shinikario looked at in-hanes and body burdens in children age three to 11 years and the endpoint was developmental. And they were looking at the height, weight and BMI in those children. And they used actually the MRLs that we have. And they just looked at the four chemicals before PFNA, PFOS and PFHXS. And again, they used the P4 as an index chemical. And they concluded that the height was actually associated with the higher mixtures index. For the boys. But the thing is that again, this is very limited just to four chemicals and body burden of those chemicals. So, I think it's a good start. You know, but we have a lot to learn about those chemicals because they are 5,000 or something like that around. And it's not that easy to actually go and derive those PFs or RES. But the thing is that we don't have a PF also for all the dioxins. We just use the PFs for the dioxins that are most likely to be in human blood. And that's maybe the thing to look at, you know, just look at what is the body burden in humans and at least make a risk assessment for these chemicals. That's it. I'm going to add, Elsie. The, Scott, that was a really good question. But the real problem with PFAS is they are not like PHAs or PCBs, which are group of chemicals which have a nucleus, a main nucleus. And then there are structures around that nucleus. So PHAs and PCBs have a specific structure. And then the substitution around them is different. So the diversity of critical structures in PFASs are going to dictate what kind of ADME they go through. And ultimately, when they get to the receptor, there are so many different receptors they act, even though PPR alpha is the important one, but they can also act with card, they cannot with ER. So there's a bunch of receptors they can interact. And the whole idea of the TEF and relative potency factor, both of them is based on the concept that the dose response curves are congruent. That is the potency is different. They are dilution of one another, but their dose response curves are similar so that you can then add them up as a potency weighted dose addition. So we need to have a lot of data to make sure the dose response is forward at various doses so that we can do the TEF calculations. And we might have to break them down into subgroups and say, and maybe in the next session, we'll be talking about how to subgroup and all that. And that's where maybe some TEFs and relative potency factors can be calculated. My thought is really important in deriving these numbers is the mechanism of action. They all display that we need the same mechanism of action. And we don't know a lot of these mechanisms for many of the chemicals that are included in PFAS. Yeah, I agree. Okay, we have a couple of minutes left. Tom, do you wanna ask a quick final question? You're on mute, I think. Tom, you're on mute. Sorry about that. Yeah, about methods for looking at health effects associated with mixtures in epidemiology studies. We talked about it a little bit in the last session, but if anybody here would like to comment on that. Tom, that's a good question. Good seeing you actually. You know, it's interesting. Toxicologists like controlling things, controlled exposures, and epidemiologists like with natural experiments, which are happened. So there's an accident that happened. A population is being exposed to somewhere. And then we are trying to find out, I think yesterday, a couple of times this came up and in our publications, she's been way back in 90s, we wrote that the experimental scientists and the modeller and the risk assessor should get together at the start so that they can all see what each other saw and how they can see what each other's needs are and be able to do things. You know, risk assessors are in the field out there. They need to solve a problem right now and they need to come up with a solution. So we might develop all these complex models and conceptually realize all this can be done, but there have to be data to support. And I think that's what we are talking in the last few days. There are so many data gaps, but it is great. So I think that the epidemiologist and toxicologist and risk assessors can work together. And I think as we go on in the session, I would advocate that the more cooperation between federal agencies doing various things and the academic institutions doing different things. I think that would be really fruitful. And so I look forward to that, but anyone else want to add something to it. Okay, well, I think we're ready to move on to our next session at this time. The next discussion will address class-based approaches to risk assessment with the overview presentation provided by Rusty Thomas, the director for the Center for Computational Toxicology and Exposure in the US EPA Office of Research and Development. So thank you and welcome again Rusty. Thank you. And I appreciate the opportunity to address this particular topic area. The agency's engaged. This was not just an EPA effort. The agency engaged in sort of pulling us all together were at the top EPA, ATSDR and the Department of Defense. And so both contributing to the summary of the state of science as well as ongoing activities and potential data gaps. And so I think the current state of the science within the class-based approach, the group felt that there is an evolving definition of really what constitutes a PFAS. There continues to be an evolving definition of what a PFAS is, but that there's continued progress in capturing the scope and breadth of the PFAS landscape. We continue to know more about what's captured in that PFAS landscape and better defining different parts of that landscape as well. I think there's emerging consensus on the need to use class-based approaches to inform decisions on PFAS, do the number of PFAS in commerce and environment. I think that in most cases, and we've heard over the last couple of days, right, nobody, I haven't heard anybody say, no, we need to take all these PFAS one at a time. And I don't think anybody has really argued for doing that necessarily. And so I think we are all pretty well aligned with the need in order to have some class-based approach to this issue. There have been multiple class-based approaches that have been proposed based on structural considerations as well as various intrinsic properties, such as persistence, mobility, bioaccumulation, exposure and effects, right? I think many of those are under discussions and the pros and cons related to each of those proposed class-based approaches have been discussed already. I think part of what we'll do today is continuing that discussion as well. Many of these class-based approaches rely on subjective definitions though of class membership, thereby limiting consistent and reproducible applications. So you may have one person that puts a PFAS in one approach under one circumstances but puts in a different class in a different context or by a different person that's evaluating that. So I think there's a need to better define these definitions of class membership more rigorously. For human health endpoints within the EPA, PFAS analogs and or groups are typically defined based on a combination of chain length and functional group. We talked about that yesterday and that the number of PFAS analogs in our groups and the associated divisions are really dependent on the availability of toxicity data or lack thereof. And so essentially what that says is what's driving that definition and the number of those classes is really the availability of data. Current research listed here, improved understanding of the mechanistic and toxicokinetic properties and relative potencies of about 120 structurally diverse PFAS are going to be available in relatively soon. I'm using these in vitro and high throughput toxicokinetic as well as toxicity type assays. And we believe that this is going to help inform PFAS grouping as well as identify additional testing needs. And we discussed this yesterday. Also ongoing research is how to better create objective and reproducible structural groups of PFAS based on things like Markoo structural representations and chemical fingerprints. We also have going on an improved understanding of chemical identity, including molecular structures, PFAS and mixtures byproducts resulting from manufacturer processing use disposal and existing information concerning the environmental and health effects of PFAS. This is being captured under a proposed rule that was directed by the National Defense Authorization Act. So there's a proposed rule within the EPA that is going to help capture this information and report this information related to many of these PFAS that are currently being used. And there's an improved understanding of exposures to approximately 172 PFAS added to the toxic release inventory in 2020. And this is to help also inform some class-based approaches as well. Lastly, potential data gaps. And I think there's a lot of them. And even though there's only three bullets here, it's a pretty large set of data gaps. I think it's fair to say that there is really insufficient toxicological as well as toxicoconetic data to group these PFAS that are currently in commercial use and in the environment on a robust mechanistic basis. And I think we're trying to address some of that, but that's still going to, even after the previous slide, that's still going to remain. There's limited intrinsic property exposure and effects data to inform selection among the various class-based approaches that are being proposed. And the mechanistic properties of volatile, semi-volatile, and non-DMSO-soluble PFAS is also a data gap. We're really evaluating a fairly narrow physical chemical space of PFAS. And we're not addressing some of these other types of PFAS that are volatile, for example, in some of our toxicological evaluations so far. So that's going to continue to be a data gap. And I think I'll turn it back over to you. I'll see. Okay, great. Thank you, Rusty. That was a great overview of some of the data gaps that we need to fill. So now we'll bring in our additional federal agency to discuss, and David Dorman and Tom Webster from the planning committee for follow-up discussion and questions. And Tom, maybe you can kick us off with a first question. Hey, Thank you, Rusty. That was really good. So, you know, as you know, class-based approaches are being talked about based on various intrinsic properties, persistence, mobility, bio accumulation, toxicity, some sub-subset of those. And I'm wondering, you touched on this a bit. What do you think is sort of the most important research needs in order to go down that route? I'll hopefully let some of my other colleagues chime in as well. I covered a few of them, as you said, in what I had articulated as the, some of the data gaps and the ongoing approaches. I think most of the priorities, I would argue are at least beginning to be addressed by these ongoing research activities and the data gaps are provided. You know, these include those robust ways to create objective reproducible structural groups of PFAS and also more empirical data that capture these intrinsic properties across a broader range of these molecules, right? I mean, we're really at a data poor and data star state that to help inform these categories and groups, right? And those aren't necessarily to just create those structural groups, although that's still a challenge as well. But in order to be able to select from among those different class-based approaches, we're going to need data to inform that as well. Because I imagine that there's not going to be just one class-based approach for every problem. I think we've discussed in the past and in certain ways that certain decision contexts may require a slightly different class-based approach, right? And so it's not necessarily a one-size-fits-all. And I think that the solution to that is developing these robust data sets across these different intrinsic properties and other exposure and hazard and uses in order to be able to more effectively tailor your class-based approaches to your particular decision context. And then probably the last data gap that wasn't necessarily listed in those data gaps, but probably should have been is the fact that we're going to need a sufficient number of data-rich analogs in many of those, if not all of those categories or groups in order to inform those potential human health effects. And those were currently lacking as well. I'm not sure if anyone else wants to weigh in, but I have just a small follow-up question. Grace or Barry, did I miss anything? No. No, I think you covered everything that came to my mind. So I was wondering, so given the tremendous data challenges that we face right now for some of the toxicological properties and data needs in terms of some of the in vitro testing, I wonder what you would all say is the potential of these in-silico methods. So you're using chemical structure to try to fill in some of those other intrinsic properties. So things like persistence, because that seems to be the thing that we have most available right now and can mine. And when we're thinking about timeframes of data availability, that's the thing that is most at hand. What's your perspective on that? And what are your plans to sort of expand in-silico modeling for these newer PFAS? Can I ask a clarification there in terms of the in-silico tools, because I'm not really clear whether you mean sort of existing silico tools that exist for prediction of things like biocumulation or biodegradation, or are you talking about the development of new models that can address that need? Because one of the challenges I sort of see is whilst there are a plethora of different models that can predict some of those endpoints, often they're lacking the underlying data, PFAS-related substances is often lacking. So the extent to which we can make meaningful and robust predictions is a challenge. So I kind of wanted to ask the question to clarify what you were really intended there. I meant both. And so your perspectives that you're offering right now are valuable. So where are we in being able to use these in-silico tools given that that's something that we can use more immediately or develop more immediately? Race hit upon, hit the nail on the head and that in many cases are existing models aren't optimal predicting some of these key properties within the PFAS space, right? We are, although it's not necessarily listed on some of these slides, have an effort to both curate that type of information that is available in the literature and in other more gray literature that exists. So can you call all the physical chemical properties or environmental bio accumulation type properties from a broader set of a public and then use that to better train or update our existing models to make them more robust within this space. And that's probably one of the efforts that is going on that will hopefully improve our ability to do in-silico predictions within this space. We'll see how successful that is, but that's an ongoing activity as well. Sounds great. Thank you, Rusty. Dave, did you want to ask another question? Yeah, it's a little bit of a follow-up. So Rusty, one of the challenges when class-based approaches have been looked at for other classes of compounds is sometimes the in vitro and computational approaches don't really predict terribly well the outcome data available either in epidemiology or toxicology studies. And we have a problem for both the legacy compounds as well as the emerging PFAS compounds. So how do you think you're going to try to address that since it sounds like most of the effort is really trying to come up with a mechanistically-based binning approach? Yeah, I think, I mean, that's certainly been a challenge with some existing class-based approaches is that tying some of these mechanistic-related assays and information that you're getting to more of these higher-order apical effects and showing that alignment, certainly. I think the way we are trying to approach that in these class-based approach that we're taking is to use that mechanistic information within that to help, I guess, inform the existing structural groupings that we already have. I think we talked about a little bit yesterday that we have certain structural groupings that based on chain length and the functional group. And can we, in certain groups that we have already where we have a data-rich anchor chemical, do we do all those molecules that are within that group, do they all have consistent mechanistic data that within that class? Are they all activating the same set of receptors in biological pathways or do they clearly separate some of them within that particular group may activate the estrogen receptor and half of them may not? And so maybe you need to break apart that structural group that you've used traditionally and that may not necessarily be evident in some of the in vivo toxicology information that we've collected. And so trying to fold that mechanistic information in sort of a weight of evidence in order to know whether you have a class that is mechanistically consistent as well as structurally consistent based on what we've used previously. So that's kind of the approach that we're taking with this. Will we be entirely aligned with all of our epidemiological data and toxicology data? It would be nice, but I think all of us realize that's probably not going to happen 100% of the time. Thanks. Didn't anyone else want to weigh in on that class question? I have another question, if not, which is, you know, as we take a step back and think about a class-based approach to chemicals management. And this has been done for a few other classes of toxicants like the flame retardants. You know, are there lessons learned in the pathway of research toward developing those classes that could be extended to PFAS or were there steps along the research journey and filling some of these large gaps that you've outlined today that could be filled to move us in that direction? My other colleagues take first crack, if they would like. I'm not sure about that one, Rusty speaker. You know, I've kind of read about the flame retardants class approach per se. I'm very familiar with the literature in terms of sort of category-based approaches. But what I see more of late is kind of a sea change in terms of how we fold in some of this mechanistic data. And I don't think we're at a mature enough level in terms of, you know, how that can be best done in order to, you know, say, you know, it's kind of novel and cutting edge in terms of how we're trying to do that. And I don't know if there are that many lessons that we can draw from what's been done in the past that we can bring forward. There's certainly the current state of the art that's been sort of discussed that we're applying here. But I think, you know, the use of this type of data that we're using is kind of cutting edge, is novel. And, you know, I think we're kind of learning as we're doing. I think it was also, Lisa, discussed in a previous mixture related talk about how, you know, the diversity, the most structurally physical chemical property and in every way, shape or form, the PFAS are much more complicated in a bigger animal than some of those other examples that you're talking about, the, you know, the bromelated flame retardants, right? And so I think although we can take away a few of the lessons that were used to derive some of those class-based approaches, they probably only go so far when they come to the issue related to PFAS. The extrapolation there is going to be a challenge because of some of those characteristics related to PFAS that we've discussed here as well as previously. I think to build on that too, I think, you know, with some of those previous attempts, there's been a much more of a finite definition of what encompasses that landscape for some of those class-based approaches, whereas, you know, we're still wrestling in a sense and this was sort of articulated in Rusty's first slide in terms of, you know, have we reached that consensus or harmonization of what constitutes the PFAS and how to define the boundary of what that landscape is and that makes, again, a challenging exercise in terms of trying to sort of tease apart what should be the classes and sub-categorizations within that space. Great. Thank you for that. I'm going to hand it to Tom Webster to ask the next question. Yeah, Rusty, I did appreciate your statement about their different contexts. I think of at least two kinds of contexts. One is existing contamination, like we would find, say, in groundwater or something, and the second is sort of ongoing production and new products. So I'd really like to talk a little bit about the latter. For example, new PFASs and things in food contact materials and stuff like that where sometimes we may not even know what's really there or very much about the toxicity. What the heck do we do about that? From a sort of class point of view and what sort of research do we need to do to be able to think about those coherently? Yeah. Well, that's a good question. I mean, in terms of taking that, obviously there are a lot of data gaps associated with some of those novel PFASs and what they can use and then do uses and how they can be applied. Certainly that is related to some of the exposure discussions that we've had previously here and how to characterize that using some of the newer technologies like NTA and others as well as identifying what the data gaps are so that we can have a better understanding and appreciation of the exposure-related aspects of those PFASs. I think that's all contained within it. But specifically, I think you're talking about different class-based approaches towards understanding and characterizing the PFAS that could be related to some of those new uses if I understand it correctly. And I think that taking that class-based approach and applying it to your question, I think that in that class-based approach, you have obviously different physical chemical properties and structural characteristics that are amenable for application to a specific functional use. And so I think you can use, to some degree, some of those class-based characteristics that constrain its functional use as well as some of the characteristics related to exposure and other related aspects to sort of narrow down both your exposure-related properties as well as some of the potential transport and other related aspects that in order to characterize those exposure pathways and better inform some of those particular properties and questions that we may have. But I hope I answered your question from a more class-based context, right? Yeah, these are difficult questions, so I'm sorry to ask you such hard questions. No, it's just also fitting in a class-based approach to new use, sorry. Yeah, no, I mean, I think the idea behind a class-based approach is we have a tremendous number of new compounds in new materials like these food contact materials. And we're lacking a lot of the data that you outlined that we need, and certainly we're struggling with even the basic chemical-physical data for those new compounds that would allow, even in silico modeling or some kind of QSAR model. So we're just trying to get our heads around, well, what can we do about those chemicals and what might be next steps? So if you guys want to weigh in with any final thoughts on those new compounds and how, you know, if or how research is needed to inform class-based approaches, that would be great. I don't know, Barry, if you had anything you wanted to add to our discussion here. I just had one last point maybe I wanted to make, to add, especially to the newer classes of compounds, the newer manufacturing techniques are likely to have fewer, I guess you would say, by-products and things like that, which you might have seen then with the older electrofluorination techniques that might have had and might have produced more branched isomers and unintended by-products of things like PFAS and PFOA. So even, so the newer classes of PFAS might just end up being more specific, I guess you could say. And while our more legacy contaminants are probably more diverse in their structures. So that could be a consideration going forward in terms of just the variety of exposure that we're looking at if we're talking about newer materials versus say ground water. So I just wanted to add that in. Okay, that's an excellent point. Thank you very much. So thank you to all of you. I think we're going to move on and just transition now into our last session on emerging issues. We've decided to forgo an overview presentation based on this topic and focus entirely on discussion among our various federal agency discussants. And we have Gloria Post and Laurel Shader from our workshop planning committee who will also participate in this discussion. So welcome everyone for this discussion on emerging issues as we try to get our heads around all the complex information that we've heard about over the past couple of days. And I think Gloria, you wanted to kick us off with the first question. Thank you, Elsie. So as everyone knows, newly identified PFAS with very little or no chemical, physical or toxicology data are continually being detected in drinking water as well as other media and creating concerns among communities and the needs for addressing and evaluating the risks in a timely manner and providing public health protective advice. I think it's important for everyone to recognize that currently risk assessment and guidance values cannot be developed without at least a minimal in vivo data set. And this will not change in the near future. So my question is, can NTP and EPA continue their highly valuable focused in vivo PFAS research as needed to continue to support the needs of states and others in addressing public health concerns along with the development of their in vitro research program that will provide data and evaluation of a large number of PFAS. But I'll make a quick shot at this and Andy, please get after me. So, Gloria, I think one thing that may not have been clear is with EPA is that we don't just have an in vitro or high throughput characterization of these, and try to mechanistic characterization of these molecules, right? In parallel and ongoing in that is to also have some targeted in vivo studies as part of that second tier, right? And so what we're trying to do with these studies is better group them and categorize them into both mechanistic and toxicoconetic based and structural based groupings of these molecules and then identify, okay, groupings mechanistic, toxicoconetic and structural groupings do we have left that don't have a data rich analog for which we can evaluate the potential hazards and potencies of those molecules within that particular group. And so the next phase of this is really to continue to fill the in vivo data gaps of those refined structural groupings going forward. So that certainly that second step is being planned for and is being executed as we go forward. So I didn't want to get the impression that the ongoing work on the mechanistic characterization is the only, we're going to stop there. I think that next step is going to be key to enable decisions for people such as yourself. Thank you for the explanation. I have been following all of this. So I have heard about the tiered plan tiered approach, but it was probably your explanation was really useful for people who might not know, but I guess the heart of my question is, this is the research is currently under development that in vitro studies and approaches are being developed and in the near or longer future, all of that will be available and the tiered will continue. What I'm saying is that in the past, the focused in vivo, when we find something in drinking water and we need to develop a guide like a guidance value that studies that NTP and EPA that focused in vivo studies, not necessarily chronic or 90 days short ones have been extremely valuable as the basis for providing risk assessment that we can developing them to address these concerns and the approach you're describing is more of a longterm thing. I would hope that when something's really needed, it wouldn't have to go. That isn't that something could be done more quickly. That could help states and others who need it. And that was probably the heart of my question. Yeah, maybe there's a follow up question, which is what can we do fast in your perspective? And that can be used currently. Yeah. So what is available in the very near term as, as research projects from your, products from your respective agencies that can help us with some of these newer PFAS questions. Well, I'll build a little bit on what Rusty said. I mean, in the, in the near term, you know, the first thing I think we're going to be producing, which I would hope is useful is the, the results from the experiment that we discussed, the high throughput testing of roughly 120 PFAS, a variety of different end points. In addition to the grouping work, which was just discussed, you know, we will be using that information to prioritize some chemicals. We have some, some in vivo studies underway right now. For example, our researchers have collaborated with others on some studies of Gen X, HFPO diamond acid and others. As you can imagine, that's, that's much more complex and slower work. It just takes a lot more resources to do those chemical at a time. And we'll continue. I'm not going to claim that EPA is going to, you know, answer all the questions and all the chemicals, but our scientists in the in vivo work will continue to be part of the larger research community that does research in part indicated, you know, prioritized by the chemicals that we find from the high throughput work. Great. Laurel, does you want to ask the next question? Sure. Thanks. So we've talked a lot over the last few days about non-targeted analysis of the way to learn about the growing number of PFAS, but these analyses can't provide quantitative information in the same way that we can with targeted analyses and an analytical standard to calibrate the instrument. What factors limit the availability of analytical standards for new PFAS, including precursor compounds and how can this be addressed? Well, I'll talk a little bit about that. Maybe some of my other colleagues can add on. Yeah, that's certainly a challenge. I mean, when we discover these new chemicals, we often don't even know what they are, let alone know where to go for a reference. We do have some tools, finger tips, you know, we have made some exploratory efforts at working with different chemical vendors who can synthesize chemicals according to a structure that we give them. And for things that we think might be important, we can go ahead and try to generate a standard. There also have been some, you know, strategies which have been willing to share laboratory grade standards with us once we find something, for example, in a river and in water that we don't know what it is. So there's, you know, there are ways that we can get to a standard and get to better quantitation. I'm not enough of an expert analytical chemist to know whether there are some purely computational solutions to coming up with better quantitation. I suspect there are. And I know there's a lot of people working on that science and we'll probably make some progress there over time. But as we do identify things which we have an indication, we think are relatively common. We can't explore means of, if not purchasing a standard of acquiring one, synthesizing one, de novo. Laura, I think the question is one challenge. So from the bio-monetary perspective, having, you know, analytical standards is crucial. And oftentimes having to get them synthesized to pay for is not a true thing. It takes a long period of time to add a new analytical method to a suite of methods for new chemical that passes all the quality control, you know, requirements, and now stands up to something that we're going to stand behind and something like in Haines and the national data support. So while we're constantly looking to add new things and do all the steps, it's not a trivial thing to say, we want to just add new chemicals. It takes time. It takes work on the analytical canvas. And if you want to have a reference method that you can stand behind and you can make available to, you know, laboratories across the country, that's not a, that's not a trivial thing. Those are great points, Pat. I think we, you know, anyone who tries to do this kind of chemistry would agree with you. And so I guess that brings us back to this discussion. We've heard a lot over the past couple of days about the potential for non targeted analysis and high resolution mass spectrometry, identifying these new compounds and producing a lot of really interesting new information. And I guess I would like to hear more about the, the federal agency's plans to use that type of data. How do we ensure that the type of data being produced in by non targeted analysis is in fact, you know, truly reproducible. There's been some discussion in the literature on this. And then there's this issue of when we think in a cross cutting way about risk assessment, I think we want to focus first on the ones in the environment that are most abundant of those new compounds, if you would all agree. And so how do we prioritize on these suspect screening lists, which right now, you know, we're getting a ton of data, which is wonderful. But on the other hand, we're getting a ton of data. So, so how do we prioritize that information? And what are the plans from a research perspective to, to moving toward more quantitative information? Okay, I'll, I'll start on that one again. And then some of my colleagues can add to it. So to answer your last part first, you know, the non targeted method development is very much a part of our research program at EPA. And we know that some of our colleagues in other agencies do as well. You know, we use it for somewhat different purposes than what Pat just talked about. We're not necessarily aiming for a method that we could incorporate into a nationwide biomonitoring approach. Typically these methods go through a life cycle. So they start first as kind of a research method. You try different things. You find out what works. You get something that's stable. You start publishing on that. You collaborate with others. And then over time, if they're, if there is shown to be a benefit from that method, you can move towards trying to develop an SOP, standard operating procedure to develop a standard method and then put it through a single lab validation and a multi-lab validation. We've done that, for example, with our drinking water methods. Method 537 was updated. Method 533 came out recently. We're working with Department of Defense right now to do, to validate a standard method for non drinking water. We've got method development well underway for air emissions. And I expect that the non targeted methods will go through that same life cycle. So a year or two from now, we'll hopefully start working towards some kind of an SOP across the field, maybe of professionals. That being said, you know, there are still some barriers to implementing that in a nationwide kind of a survey. There's, there's expensive equipment and training, which we hope and anticipate will become more accessible over time. So it is going to be an evolution, but I think the science of high resolution mass spectrometry non targeted analysis will advance really just like all the other mass spec work has over the last 10 or 20 years. Benjamin. Yeah. I was just going to mention. One of the keys to just non target analysis in general. So this also along with EPA has a very large research program and non target analysis. And one of the keys that we've identified is really in order to start approaching a standard method or even a standard approach for non target analysis. We first need to get everyone on the same playing field when it comes to reporting performance metrics and how do we describe our data. And so there is actually ongoing efforts. There's on our efforts other non targeted omic style approaches, but it's right now we're in the progress of developing these standards. But there is an effort right now. It's an international working group that EPA members and I'm a part of as well that has been working on trying to develop reference ways to report your standard, your results and come up with ways to harmonize methods. And I think that that's going to be key to first being able to understand your performance of your method. But then getting to your question about how can we validate these methods and how can we better use our information is we need to understand the performance to understand what the uncertainty is about our identifications. So there are previous papers. There's previous research about reporting confidence. But I think we need to get further down the road to understanding that when I identify compound, this is much competence I have in that. And then our hope is that once we get to there, we can kind of start going down the road with things like semi-quantation. There's a very large push in the community to provide, even if we're not going to be able to provide solid quantitative numbers, provide some sort of okay number to give us a better idea of if there's a lot or a little and something like drinking water. And that will help us be able to prioritize better as if we can get at least a general idea of what the kind of concentrations we're looking at is something that's part of the general non-targeted community, but also part of this PFAS non-targeted community. Yeah, those are great points. Thank you for that. So if I can get my head around this correctly, then there's a lot of progress being made in, in non-targeted analysis and there's a lot of support among the federal agencies for developing those SOPs. And there's some thought being given to semi-quantitative interpretation of those results. So how about then on the methods that are also out in the academic literature right now, what's the federal agency thinking on those? So there are various total flooring methods. They were mentioned yesterday and you guys discussed them a little. I sensed and perhaps incorrectly a little less enthusiasm for those methods, but those methods, you know, like the top assay, combustion, ion chromatography, the, the piggy analysis from Grand Piesli's group, you know, there's a whole bunch of different methods emerging. Perhaps you guys could comment on some of the challenges in developing a standardized protocol for those methods. And if, you know, what, what the federal agency, or how the federal agencies view those total flooring or total organic flooring methods moving forward. Okay. I'll kick us off again. And then my other colleagues can join in. Yes. All the, all the things you mentioned are, you know, on our radar screen, we're doing research on some of them. We're working on a tough method both for, for water as well as for, for air emissions. We've used the top method before. I mean, I kind of view these things, all those tools in the toolbox and there, there's some of them are very relevant for certain kinds of applications. Some of them may be broader. I like to think that we could reach a place with top, for example, where you could use it as a screening test. So if it could be done in such a way that it was not terribly expensive, then, then localities could use that to screen samples first for presence or absence, kind of like a litmus test and then do a deeper dive when they find tough, but all the things you mentioned, the piggy, the CIC, these are things that either we're working on or we're funding in some cases, extramural projects through our star grants to work on. You know, we, we want to move very much towards some kind of a toolbox and that's not to say that every method has to be, has to go through all the steps that you do to get to a standard validated method. For example, for a regulatory purpose, some of these methods could be very useful if they're simply published in QAID. We're in the process of building a website right now for our methods and we're going to be adding, for example, our publications and SOPs, you know, for some of the research grade methods as we call them. So I think everything you mentioned and more, as more equipment, more approaches become available, will be useful, but they'll probably be useful in different ways. I don't know that we plan to take everything to the same kind of multi-lab validated level, nor do we need to. Anyone else want to add to that, Benjamin? No. No, I was, I was just going to say one of the bigger aspects that we look at, especially with NIST and talking about reference materials when it comes to these overall total organic fluorine or the top assay is, is the need for standards, but also a mechanistic understanding of what's going on so that you can do things like close the mass balance. And so we don't necessarily explore that because based on a lot of our materials are, are targeted materials for, for our SRMs. And so we don't necessarily explore this whore holistic approach. It's on our radar, but we haven't, it hasn't been something we've explored right now. Okay, great. So Gloria, I wanted to hand it back to you to ask a different question. And you're on mute right now. So if you could just unmute yourself. Thanks. Thank you. How can we develop more comprehensive publicly available data on chemical production use and discharge, particularly for those PFAS that have limited physical, chemical or toxicity data. What data are publicly available and what data are current, what data are not available or not publicly available? And how do these data gaps affect our understanding of risk? Relevant to these questions. Data that has been provided to EPA. As confidential business information. By industries that use or make the PFAS may not be available, even to federal researchers. And this requires that the federal research. And also that the research and distribution of these resources be used to conduct a sort of detective work to figure out information that may already exist. Is it appropriate to find ways to decide which of this information might be appropriate to make publicly available and to do so. If appropriate. All right, I'll start on that. And then some other colleagues can also chime in. You know, we're doing some research and we're doing some research and we're doing some research on some of the things that are being done for a kind of things, which are just outside the domain of the science that I do, that we do in the research part of EPA, and which really are kind of the scope of this meeting. So we, we basically work within the law as it's written right now. And, and we're, we're content to do the detective work to apply our non targeted methods to try to find out as much as we can about the chemicals. That are in the environment. And then to the extent that we can flag those for the, for the, for the better, for the better. And that's a good question. I think you answered it well, Andy, that. We do have to live within the constraints of the law and what information is available. Okay. I want to thank everyone for participating in this emerging issues session. And I think what we're going to do now is. We're going to bring back our planning committee participants for questions to join the last panel and ask the other federal agency representatives who wish to answer a question. To raise their hand. Thank. And, and we'll just open it up now to, to questions from the planning committee. We're full of questions after two days of information. So if everyone wants to take a breath and clear their heads. And again, we want to think about from a, from a. Cross cutting issues and risk assessment, given all that we've, we've learned. You know, what, what are the big questions that remain. In our head. And maybe Dave, you could kick us off with a. A first question. Yeah. So this kind of crosses both the epi and the human. Data. As well as the animal data and kind of is clearly cross cutting, which is comes down to dose response and the disconnect between high dose studies, say conducted in animals and concentrations seen in environmental studies. So how do we bridge that? I mean, it's that issues. Common to lots of problems in toxicology, but it seems to be magnified for people. And I think we're going to, we're going to come into lots of problems in toxicology, but it seems to be magnified for PFAS because one thing I got struck with yesterday was the feeling of chemists trying to, what I'll almost say is like chasing zero, so to speak, like how low do we need to go in order to be able to be confident about human exposures. And then how do we merge that with the toxicology data and mechanistic data? I would just make one comment on that. At least with our experience with the animal studies, and this kind of relates to a point I brought up yesterday about the importance of quantifying internal exposures and particularly serum or tissue concentrations is that the higher animal doses based on species differences and clearance rates can result in serum concentrations that are not wildly discordant with human exposures. We, our group has a Gen X paper that's actually published online today in environment international on the developmental talks toxicity of Gen X and a lab rat model. And in that we have one figure that shows how the maternal serum Gen X concentrations relate to serum concentrations from fluorochemical workers. So that would be more of a potentially worst case scenario than the general public, but I think it kind of reiterates the point of the importance of characterizing the internal dose as compared to the external dose because sometimes that can seem out of step with what people may be experiencing. Thank you. Any other federal partners want to weigh in on that? Well, I would just add a little bit to what Justin said. You know, part of part of David's question was having to do with the ability to measure, I think very, very minor quantities. And that is true. And often those are for different purposes. I mean, we want to measure things so that we can discover them. And so we can find out what's in the environment. I don't, I would not say necessarily everything that we find is at a level that's going to cause some kind of adverse effect immediately. But I think if we can find those places that have much higher doses that people are being exposed to, that much higher risk, I think that's part of the value of the application of the methods. The fact that we can go down to a very low level is informative because it just helps us, I think, find out quicker where those higher exposures might be. And maybe that segues nicely into a question that was posted by Mark Johnson, which is, have the federal agencies come together and ranking those PFAS of highest importance? And perhaps that could help in prioritizing how some of the data gaps are filled. So I'm not aware that we've collectively sat down and tried to prioritize, you know, from one to 100, a list of the top PFAS. We are maintaining awareness of each other's work. And we're all, you know, most of the work takes place with a similar set of 20 or 30 PFAS that have been well-established and that we know are out there. I think we're going to be adding dramatically to that list over time as we apply more of these non-targeted methods. And we'll probably find that there isn't any single ranking of chemicals. It's going to depend very much on where you are in the country and what it is you're interested in, whether you're interested in remediation and treatment and cleanup or whether you're interested in toxicology. So I don't know that there's a single list, but I know that we have a mechanism in place through our cross-federal technical working group that enables us to share that information and to keep each other aware of what we're finding and what we personally or our agencies think of as important. I don't know if any of my TWG colleagues want to also... I'm going to echo that. You know, so we know exactly what chemicals are being tested in NTP and META, what chemicals they measure. And we tailor in many ways our measurements to what's being measured out in environments. We want to target the chemicals that are most common occurring in environmental settings, particularly environmental settings where people explode. And then for our health studies, we're banking by a lot of samples. So in the future, if more chemicals come out of that concern list, we'll be able to go back and look at them and going retrospectively as well. So we're trying to be facile enough just to look at the ones that we know are out there now, the only ones that have good measurement and nothing's wrong. So I admit that in some ways, we're looking for the light and the lamp posts, the car keys and the lamp posts in that regard. But we're also worried about what you might know about in the future. We're trying to preserve the ability to investigate those things in the future. And we do look more carefully at the EPA and follow the methods they're developing what they're measuring and tailor our work based on that. Okay, great. It's great to hear about the coordinated efforts. Does anyone else want to weigh in on that before we move to the next question? If not, maybe. I have one minor comment related to that in terms of our laboratory work. Sometimes, you know, we'll see a biomonitoring study. For example, the one NC States doing here in North Carolina that will report frequent and or high serum concentrations for some novel or emerging performative compounds. And from an analytical chemical perspective, they're able to do those detections and using relatively small masses of a maybe a laboratory purified version of the compound as compared to then us trying to translate that into an animal study. And many of these are oftentimes we find not commercially available in a purified standard form for us to be able to purchase and run studies with. So I would say that that's one limitation, at least that we have experienced is access to some of the more novel and data core compounds that are detected but not available in the amount needed to do some hypothesis driven experimental work. Have you ever just a quick follow up? Has that ever been discussed with some of your other federal partners? I was just thinking back to Andy's description of industry collaborating to provide that kind of information or their ability to synthesize standards quickly. So has that kind of collaboration ever taken place to push the research forward or is that something that could perhaps be built on? We have discussed that with chemists like Mark Stryner and James McCord here at EPA in terms of if we're interested in something conferring with them if they know of a way to get more and then oftentimes that can sometimes lead to oh, there's a university researcher who is an organic chemist that can synthesize it but oftentimes you're still limited by the amount that somebody can produce under non kind of industrial commercial processes. You're right, there could be some opportunity there potentially for a cross agency. At least I have an answer as well to that and one of the things that I highlighted in some of our presentations is we've assembled this library of 430 PFAS that were commercially available and made that available across all of our federal agency partners. These are fairly structurally diverse but these are the ones that were commercially procurable at a reasonable analytical purity and so those are all made available more broadly for hypothesis driven or experimental work or even to allow for the refinement of methods and so we use that in multiple different ways that library in different ways in order to facilitate that. That's great to hear. Thank you all for those perspectives. I'm going to pass it to Tom Webster to ask another question. Hey, when I think about the mixtures and how if we want to go sort of a mechanistic group we need to understand what the end point is and so that makes me think about it might differ on the end point and so we might want to focus on end points that we think are maybe the most sensitive at lowest doses and have the biggest public health impact and so it's kind of a big picture question about that like of all the end points that we've looked at which ones do you think are the most important but with respect to going into future about mixtures and class approaches. Anybody has any thoughts about that? Tom, this is Moiz Montaz. Can you hear me? Yeah, we can hear you. I think it's a important question and as we have discussed earlier that the PFAS is for example the toxicological profile at ESDR drafted in 2018 has P4, P4, PF and the hexane sulfonate. There are data for those four where we could do a detailed analysis and derive MRLs for them. If you look at the receptor level they interact with PPAR alpha they can react with CAR, they can react with ER, alpha and LXR. So liver endocrine disruption carcinogenesis and other effects. So there is there could be a crosstalk going on between all these receptors and all the data we are generating and Rusty's group is generating all those have to be tied together and when we do the AOP it's not linear. There's different levels so there's a whole network going on and I think toxicology has to move more towards a systems approach rather than focusing on one endpoint. It is good to focus on an endpoint but ultimately they are all interconnected and when we were doing our PhDs at Tom in 80s we were looking at hepatotoxicity or renal toxicity and just focusing on that but we know there are other effects that chemical is causing but we were not interested we just took the liver out and we made the microsomes and see what's the difference and look at the hepatocytes for the difference we never cared about what's happening over there in the brain. So I think we have to take a fast forward overview of how we tie all these things together you know the whole systems biology some kind of informatics and learning machine all these have to be tied together to develop a model that people can use but again when we develop such a model which is we can conceptualize and even practically tie the equations together to run a model like a pvk ultimately we'll have to develop a guidance so that we can a risk assessor can use that for practical purposes because ultimately in the field people don't have all these computers and analysis power to do so I think we'll ultimately have to bring it down to a simple and usable tool but I think systems approach is where we have to go and all these little asses which we are doing in the high throughput have to be tied together so that we can understand what is going on the whole system I would agree I mean I think we need some simple approaches now for example maybe the best we can do right now is just add stuff together I don't know but meanwhile we need to be doing research on this systems biology level absolutely I agree I think the HII approach is the only one which we can use and I think all the agencies from EPA to everyone New Jersey department all the health guidance values which we are developing they are all about putting them together but you are absolutely right we do need something which we can use today and those additivity has a mix approach which is the only one which we can use right now I agree with Moise but I think you were really talking about the toxic kinetics more than anything else or toxic dynamics and when you think about some of the in silico models and why they don't work too well for PFAS I think because a big descriptor a primary descriptor in a lot of these models is lipophilicity or log KOW and log KOW doesn't really work well with these ones but it does work well with these complex surfactants right I'm wondering about support we could potentially gain from maybe quantum mechanical chemist to begin to look at other aspects of the molecule to kind of help relate that to the toxicity data that we have so it's like dipole moment hydrogen bond formation things that we haven't traditionally used as descriptors maybe there's an opportunity here to look at some new descriptors and helping us improve how some of these in silico models may work on the kinetics side to help mesh with the dynamic model that Moise was talking about completely these are very funky molecules and the models don't work very well risk assessment is we can give you an answer the question is how accurate you want that answer to be and how certain you want to be how confident you want to be to apply that to communities or advice people or tell workers you are safe I think it all depends on how how acute the problem is and how quickly you want a solution the more data we have the more weight of evidence the more confident we feel Thank you for those great perspectives I think we're going to wrap up this session with just two final questions and the first will be from Gloria and the second from Laurel and I think in the interest of time we want to keep both questions fairly brief and ask the federal participants to respond with some brevity Gloria please Thanks so my question is about firefighting foam PFAS and firefighting foam because as discussed this is the most prevalent source of elevated exposure in across the country it's been discussed that AFFF contains a large number of PFAS and there's a lot of variation between batches and between old types and new types as far as the composition was reported 40 previously unknown classes of PFAS and AFFF so most of the information in the environmental data the groundwater and drinking water data and the biomonitoring data related to that are listed for a very limited number of relatively limited number of PFAS I have two questions one is research plan to address these data gaps to look at the whole suite of what's in drinking water and blood serum from AFFF exposure and also I have a question about epidemiology studies at these sites especially such as like the multi-site study where it's planned to combine the data from the seven sites to make a very large epidemiology study as I understand it could there be different exposures to unknown PFAS from firefighting foam at different sites that could affect the health endpoints and the associations and is that a potential issue in combining the data from the seven studies so there's like two questions we have two minutes remaining in our discussion so I'm sorry we can have a quick answer from the federal agencies and then maybe Laurel we can we'll see where we are with time so please Pat do you want to respond to the second one about this the studies there's a quick answer to that other than the short answer would be say yes it's possible and yes we're worried about it and you know we're defining exposure I want to remind you multiple ways and you know in any health study you can define exposure using qualitative semi-quantitative quantitative methods right and so we'll be able to look at some of the non less quantitative methods that might capture that information as well so we'll be able to look at things like years that you drank the water okay so there could be there could be a variety of ways so general exposure we're you know we are concerned about it and we will explore it as best we can and we will archive samples as I said before that should there be more information about so these other compounds that might be in the firefighting problem you identified a big issue because in fact it's not homogeneous it could be very heterogeneous and it's heterogeneous you know over time and space so those are important issues thanks okay well this has been an excellent discussion and I want to just take a moment to thank all of our speakers and discussion participants we'll now take a break until until 3 15 at which point we'll come back and go through our workshop synthesis so thank you very much everyone and we'll see you shortly