The impact of the gut on one’s overall health is quite remarkable. Knowing this leads to a very tempting proposition – map the world of bacteria in your gut and then treat imbalances to prevent and reverse disease. Sounds reasonable, but there are some major errors with this assumption. Understanding what to do and what not to do can prevent hundreds of dollars in useless testing and treatment. Today we speak with one of the leading researchers in the world in this area, professor Rob Knight.
Dr. R’s Fast Facts
Rob Knight is possibly the leading researcher in microbiota mapping.
Using microbiota mapping in clinical practice
- It’s all research at this stage, not clinical.
- Direct-to-consumer lab technology is often not the right technology for clinical use.
Thoughts on numerous labs popping up offering direct-to-consumer microbiota mapping
- Better of the labs
- CLIA certification: means you have a defined procedure that will give you the same read out consistently on the same sample, does not mean it’s accurate.
State of play on obesity and the microbiome
- In 2014, they proved they can modify the weight of a mouse via the microbiome.
- Able to show that they can predict which patients in a weight loss trial were going to lose or gain more or less weight based on the microbiome at the start of the trial, that showed that the microbiome could be predicted for weight gain or weight loss, not just response to it.
- There is causality in animal models, but not in humans.
Microbiome mapping might be more accurate in predicting relapse than calprotectin
Dynamics of the human gut microbiome in inflammatory bowel disease
- Tracked patients with ileal Crohn’s disease, colonic Crohn’s disease, and colitis.
- How the microbiome changed over time was completely different in each of these groups.
- In this study, they were able to model a healthy microbiome, define a healthy plane, then measure the distance between each individual and their healthy plane. They saw that distance and the variation of that distance over time was more predictive of their disease state than a calprotectin test (the gold standard for measuring inflammatory state).
- In order to utilize this information for clinical use, it must be validated in other populations and FDA approved clinical trials.
Does MB testing predict probiotics response?
- Microbial insufficiencies cannot be diagnosed with GI mapping.
- A whole systems approach is a better way to go, recent research has shown that mindfulness and starting an exercise program can have a positive effect on the microbiome.
Ways people can help move the science forward?
In This Episode
Prelude … 00:00:39
Episode Intro … 00:01:40
Using Microbiota Mapping In Clinical Practice … 00:03:21
CLIA Certification … 00:11:30
Map Tests … 00:13:30
Obesity And The Microbiome … 00:17:08
Dynamics Of The Human Gut Microbiome In Inflammatory Bowel Disease … 21:08
Microbial Insufficiencies Cannot Be Diagnosed With GI Mapping … 00:24:39
Episode Wrap-up … 00:26:50
Subscribe for future episodes
Dr. Michael Ruscio: Hey, everyone. Welcome to Dr. Ruscio Radio. This is Dr. Ruscio. Today, I am here with Professor Rob Knight, kind of the guy, I guess you could say one of the leading experts in the world on the microbiota, microbiota mapping. Really sharp guy, and he’s done the podcast before and he’s back today. So, Rob, thank you again for coming back.
Rob Knight: Thanks for having me.
DrMR: Absolutely a pleasure, always great talking. We had some time to connect at the symposium on natural medicine in Australia, or I guess it’s the national congress on natural medicine that was in Australia several months ago. And it was a very interesting opportunity to hear some updates from the work that you’ve been doing and very curious to get into that. You’ve been on the show before, so people should know about you.
But can you give people a very, very brief kind of background on you and your involvement in the field before we jump in?
RK: Yeah, sure. So, I’m a professor of Pediatrics and Computer Science and Engineering here at UC San Diego. I direct the Center for Microbiome Innovation here, and then I co-founded the Earth Microbiome Project and the American Gut Project. I was also involved in the Human Microbiome Project in several capacities when that was running, and many others. So my lab developed a lot of tools that are now widely used for reading out the microbiome and for analyzing it, including lab protocols and software protocols so that we understand all the complexity of that microbiome data.
Using Microbiota Mapping in Clinical Practice
DrMR: There are a few specifics regarding testing that I want to get in to in a moment, but I’d like to start us off with something that I think will be a very simple kind of question and answer but could be very helpful, and that is the use of microbiota mapping in clinical practice. And I’ll use the term microbiota mapping essentially for these tests to give you a read of all your bacteria and maybe they’re broken down into groups like phylotypes. There seems to be a lot of confusion.
As a clinician, I have lots of patients coming in who have done some type of direct-to-consumer microbiota mapping test, and they bring it in expecting to say, “Okay, I have this chronic bloating, this chronic food reactivity, this chronic brain fog. And here is my X-microbiota mapping test results. This will tell us what to do, right?” And unfortunately that’s not really what the literature seems to show.
And so, I’m curious to get your take on that. And this is both for not only consumers, but also for clinicians, because I also have clinicians who will ask me, “We did a microbiota mapping assay. And should we be treating these results directly?” I don’t think everyone’s fully aware that this is really kind of preclinical.
There are some things that we’ll talk about a little later that are really getting close. But my question is, what are you thoughts on using microbiota mapping in a clinical setting?
RK: Yeah. Right now, it’s pretty much a research rather than clinical. And although it’s good that some of those research projects have been able to help patients who were enrolled in the appropriate IRB approved protocol, in general, it’s very early stage.
I’ll just address this in a couple of contexts. So one was the American Gut Project, which I run, that’s a citizen science project where anyone can find out about their microbiome. And essentially, I go with that. We launched that as kind of a reaction to the Human Microbiome Project, because in the Human Microbiome Project, the goal was to look at a cohort of healthy people and the inclusion criteria were really rigorous, almost everyone who wanted to participate was excluded by those exclusion criteria.
And so, we wanted to do a citizen science where anyone could find out about their microbiome and where we can find out what kinds of microbiomes were out there. That’s been remarkably successful, right. We’ve processed samples from over 10,000 people at this point and we now know what a lot of microbiomes look like. But that’s not a clinical project. That’s a research project. And it’s a lot like National Geographic’s Genographic Project, right? Whereas citizen science you get to find out about your ancestry, but you simply would not want to use the results of that project to diagnose a disease. And with American Gut, it’s the same way.
So the problem is that a lot of companies are trying to use the same sort of technology, which, by the way, if we were doing a clinical project, we would use completely different technology from what we do in American Gut. There are a lot of companies trying to use the same set of technology for direct-to-consumer testing. The problem with that is in the context for research project.
So suppose you have a well-defined cohort of 1,500 people, and then you have one individual sample and you want to classify that sample according to the current disease state or prognosis based on the microbiome. You can do that pretty well, as long as all those samples are processed exactly the same way. But the different methods that different research labs use and the different methods that different companies use, they don’t provide data that are comparable at all.
And so, you kind of come in with a single profile and say, ‘can you tell me about this?’ What you need to do is you need to compare it to reference data of hundreds or thousands of other people, where you know both the clinical phenotype very well and you also know that everything was processed through a consistent method. So that’s why it’s really hard.
DrMR: Yeah, that’s kind of a good lead in, I guess, to my next question, which is, you see these different companies popping up, and just this morning I saw another one of these companies popping up. And I’m going to give people the benefit of the doubt that they probably are well-intentioned. They’re probably doing whatever they’re doing with good intent. But you read things. ‘What if disease was no longer possible? What if you could have the body you’ve always wanted? And then click here $300, $400 to get this test.’
And it’s really disheartening to me, because I know people are buying into that. And, hey, if health is a hobby for you and you have the money to dump into it and it’s not a big deal to you, okay, maybe we can make a case for that. But if you’re someone that doesn’t have a ton of money to throw at their health and is trying to get well, that’s where it really becomes disheartening to me, seeing people that are unfortunately falling victim to what are marketing claims masquerading like they’re scientifically validated, but really not.
And so, the next question I was going to ask you is what are your thoughts on these numerous direct-to-consumer labs that are popping up. And it sounds like you would not necessarily advise people on using these for clinical purposes. But is there anything that you would offer for either a clinician or a patient? They’re scrolling through Facebook, they see a thing pop up for a ‘click here to buy this lab and we’ll tell you about all your gut bacteria and then we’ll help you get healthy.’ Anything else you would offer these people in terms of what to think about that and do it, don’t do it, or whatever?
RK: Right. Well, for a particular technology, you want to look at where it’s coming from. And the most scientific literature there is that supports that technology, some technologies have been spun out of labs that are very well respected and they’re doing exactly the same thing that the professors are doing in their academic lab.
And so, if you look at something like DayTwo, which is coming out of Eran Segal and Eran Elinav out of Weizmann Institute, basically, they published a paper in Cell, which is one of the leading scientific journals, that shows that they can tell a lot about individual glucose response from the microbiome. And then there’s a company that is based on that that’s using the same technology and expanding it to other populations and that seems pretty plausible.
On the other hand, there are a lot of companies that are starting up that have people with no prior technical experience in the microbiome. That’s not necessarily bad and there’s always a room for new viewpoints and new interest. But especially when people don’t know the history of what’s been tried before and the last they’ve heard of research, very frequently there are things that we did 10 years ago that were blind alleys that we kind of abandoned. And a lot of those blind alleys are being rediscovered, and it’s a lot easier to say that you have a great technology than to actually have a great technology.
Yeah. So, you want to look at companies that are differentiated based on technical capabilities rather than based on marketing, especially right now when there’s a lot that still needs to be worked out about standards for readout and so on.
The other thing that’s really not well understood even by many clinicians unfortunately is what CLIA certification means. So, remember that all CLIA certification means is you have a defined procedure that will give you the same readouts consistently on the same samples. But that doesn’t mean that you can use that readout to diagnose any disease, right. And so I realize it’s very frustrating when you look at the awesome stuff we can do in the lab in the context of a research study and how far that is from what’s available in an FDA-approved test.
But the reason that that approval process takes so long is the FDA is very concerned about that issue you raised on people spending their scarce money that they have to spend on healthcare. You want to make sure that you maximize enough that you spend on things, so they got to be safe and effective rather than things that sound good.
Hey, guys. If you’re looking for a breath test, I’ve got some great news. You can get 20% off any testing or products at BreathTests.com when using the code RUSCIO20. That’s R-U-S-C-I-O-20. BreathTests.com is a website for Quintron. Nearly all major labs that offer breath tests are actually using Quintron products. And this is for good reason.
In 2017, Global Health and Pharma awarded Quintron best in hydrogen and methane breath testing solutions. Quintron has been used in over 100 published research studies across the world, which is why they are often referred to as the gold standard in breath testing. If you’re looking for one of the best labs for SIBO or other breath testing, Quintron is it. If you’re a clinician and you’re ordering labs for SIBO or carb malabsorption, you might want to purchase an analyzer so you can offer testing in-office. And that 20% discount is very significant here. So visit BreathTests.com. Use the code RUSCIO20 for 20% off anything, including SIBO breath testing. Okay, back to the show.
So, regarding the CLIA certification, this is essentially the clinical pathology lab board essentially—I forgot what the acronym stands for. But all labs have to be certified and approved by CLIA before they can operate legally, correct?
RK: For clinical use, yes.
DrMR: Clinical use.
RK: That’s correct, yeah.
DrMR: Are you saying that the better of the microbiota mapping labs will be CLIA certified? Is that what you mean?
RK: No, that’s not what I mean. What I mean is that a CLIA-certified test, all that means is that you can read out some technical variable. It doesn’t mean that that technical variable, once it’s read out, will help you diagnose a disease. And so, you want to be careful about that distinction, because CLIA certification is a lot easier to obtain than FDA approval for a test. The reason for that is that if the test is FDA approved, what that means is that it’s been certified by the FDA as effective for diagnosing a disease.
DrMR: That’s okay. That’s what I thought you meant. I just want to make sure I had that correctly. Now, you mentioned DayTwo is one lab. And again, and correct me if I’m wrong in this, but I just want to reiterate to people that these are maybe the better of the labs, but they still may not really have much in the way of clinical utility just yet. Is that correct? Even the better of the labs aren’t super clinically useful?
RK: Well, remember that DayTwo—so I suggest a clarification. I’m not endorsing any particular company or lab. But what’s important to remember is—so DayTwo, for example, is primarily making you buy around personalized nutrition. And it’s possible that they’ll be able to extend that into doing tests for disease states, enough to diagnose where you can modify disease.
But that’s not currently what it’s for. Like with any other thing that’s being sold for consumer pay-for-self from entertainment vehicles, using it for clinical vehicles, it’s always up to the physician’s own judgment, of course. It can be much less likely to succeed than if you have something that’s been validated directly for medical use.
DrMR: Got you. What about the GA-map test? Have you heard about this test?
RK: I don’t know that one. As you mentioned, there are just so many coming out at the moment.
DrMR: Right. Sure, sure. Okay, so something to be cautious with, and my position on this has always been somewhat clear, hopefully, I think, which is, it’s not a bad idea to support these tests for research purposes. But thinking your chronic IBS is going to have the cure uncovered by doing one of these tests is premature. And you would do well to work with a clinician and save your money in more clinically tried and true tests and treatments.
RK: IBS is an interesting case, because for years and years and years at the microbiome meetings, people were coming along with these very depressing questions on IBS, where they took IBS cases and controls and map their microbiome and saw absolutely no difference between the cases and controls. So, we did a more detailed study with the Spanish group stratifying IBS as one single dimension, so asking if IBS is very heterogeneous.
So basically, the question was are we talking about constipation or diarrhea? Are we talking about pain as mentioned by the balloon test, or no pain? And are we talking about a strong or a weak response to diet? And when you stratify people along those three axes, you see totally different microbiomes associated with those different IBS subtypes. But it all cancels out if you just treat them all with IBS. You can’t see the difference compared to controls.
So I think for a lot of these things, especially finding out what the microbiome basis of disease is, right now it’s something that you can do in a research study where you’re enrolling your cohorts and treating them all the same way. And then what we’ve seen in the American Gut is if you keep using the same protocol, you can integrate the data from new people that you have into that same population using that same protocol. But what you can’t do is you can’t use a totally different sequencing technology or DNA extraction technique or data processing technique.
So how you analyze data, computationally, has its use impacting the results. You can’t take different methods and still expect to be able to integrate it. So when you have a patient coming in with a profile and you’re trying to match that up to a research study that you read, unless the profile is generated with exactly the same techniques used in the research study, it’s essentially impossible to interpret that in terms of what it says in the research study about those particular organisms and their effect on disease.
DrMR: And thank you for saying that, because that’s something else that’s often frustrated me, where you read the specifics of the paper published by a research group, and then someone brings in a direct-to-consumer lab test that’s not using the same technology. I know this because I’ve reached out to see if some of the technology used at some of these research centers is available and it’s not available anywhere.
So you can’t take—and unfortunately, this is how it’s marketed. Maybe there’s an association to a certain skewing of the microbiota profile and being overweight. Not only is that only an association, not even a cause, so we can argue that we can’t even begin to contemplate treatment if we haven’t established causality. But the tests that Sarah Smith is bringing in is a completely different test, as you’re mentioning. So it’s for all these reasons to our audience that it’s not that I’m a hater on this testing, it’s just not there yet.
Obesity and the Microbiome
RK: Yeah, exactly. It’s very valuable in a research context. Let me tell what was stated recently on obesity and the microbiome. So, back in 2005, Jeff Gordon’s lab at Wash U, we were able to show that the microbiome was different in obese mice. Then in 2006, they showed those patients undergoing weight loss, the microbiome changed during the time of weight loss, and it was relatively consistent with the changes we saw on the mice.
Then in 2008, we were able to show that some of the same changes occur in humans. And in fact, we showed with subsequent machine learning technology, we could separate the lean from the obese individuals with 90% accuracy based on their microbiome. But on the other hand, that’s not very useful to test for obesity, right? Because you can probably tell if your patient is obese or lean without doing any microbiome sequencing, right?
But on the other hand, it’s very suggestive. And if you try to do the same thing based on human genotyping, you can only classify someone who is lean or obese with about 57% accuracy based on their SNPs, versus 90% accuracy based on their microbiome. So the microbiome is a lot more associated with human genome.
But then we did a whole bunch of work transferring the microbiome into mice. Figuring that we could do it with the human microbiome, that we could freeze that microbiome, so we could do it from a stool sample, but that would have been frozen and sent through the mail that kind of thing. And then on 2013, there’s more collaboration with Jeff Gordon’s group, we were able to show that if you take the microbiome through lean and obese humans and then transplant them in to mice, sure enough, the mice get fatter or thinner depending on whose microbiome they got.
So, that’s how you get to causality in an animal model. And then in 2014, with Tim Spector and Ruth Ley, we were able to show TwinsUK, there were some microbes that were heritable that were associated with leanness. And then if you added those microbes to that obesogenic community from an obese human and put that in the mouse, you could slim down the mouse by transferring those microbes that were associated with leanness in humans.
So, we have proved that we can modify the weight of a mouse by the microbiome. And so the causality in a mammal has been shown a benchmark for obese group. I think this was either in 2013 or 2014. They were able to show that they could predict which patients in a weight loss trial were going to lose or gain more or less weight based on the microbiome at the start of the trial. So that showed that the microbiome can predictive for weight gain or weight loss, not just responsive to it.
So, that’s current state of play. We have causality in animal models. We have predictive power in humans. We don’t causality in humans yet, but it’s relatively likely based on the other information we have.
Now, does that mean you should run out and test yourself for obesity based on your microbiome? Well, I think if your microbiome says that you’re lean but your mirror says otherwise, you should probably not believe microbiome testing, right? And obesity might seem like it’s the only example in this respect, but there’s a whole lot of other diseases that are way harder to diagnose. Exactly the same principles apply to these other diseases.
Dr. Ruscio’s Resources
Hey, everyone, in case you’re someone who is in need of help or would like to learn more, I just wanted to take a moment to let you know what resources are available. For those who would like to become a patient, you can find all that information at drruscio.com/gethelp.
For those who are looking for more of a self-help approach and/or to learn more about the gut and the microbiota, you can request to be notified when my print book becomes available at drruscio.com/gutbook. You can also get a copy of my free 25-page gut health eBook there.
And finally, if you’re a healthcare practitioner looking to learn more about my functional medicine approach, you can visit drruscio.com/review. All of these pages are at the drruscio.com URL, which is D-R-R-U-S-C-I-O dot com, then slash either ‘gethelp,’ ‘gutbook,’ or ‘review.’ Okay, back to the show
It seems that we’re getting close from what you had mentioned in Australia—and maybe things have changed since then, please let me know, but you were saying that we were getting to a point where certain microbiota results may be more predictive of IBD relapses than calprotectin, but we weren’t quite there for this to be used in the clinical setting yet, but it sounded like this may have been one of the closest to clinical application markers or items. So, can you expand on that point at all?
RK: Yeah, absolutely. That gets to another really important point, but a lot of people think that a single time point for microbiome testing is going to be completely fine. So what you’re referring to is a paper that we published earlier this year with Janet Jansson and a number of other collaborations in nature microbiology. And what we looked at is we tracked the microbiome of patients over time for ileal Crohn’s disease without resection, colonic Crohn’s disease, and ulcerative colitis.
And what was amazing about that was how the microbiomes changed over time was completely different to these different groups. And in the context of this one research study, we were able to model both the microbiome as confined to a very small region of the space that we could define like a healthy plane. And then we could measure the distance between each individual and that healthy plane. And what we saw is that distance and the variation of that distance over time was more predictive of their disease state than a calprotectin assay, which is currently the gold standard for measuring inflammatory states from a stool sample. So, that’s incredibly exciting, right?
Now, that’s in the context of a research study and some Swedish population, where you might get different results if you try it in the US population or if you tried it elsewhere around the world. So, what needs to happen to go from that study into a test that you could use clinically is we need to validate in a number of other populations that we see the same thing. We need to be able to show that either from one sample or a series of samples over time for a new subject, can we integrate them into that data frame as a reference and can we use the same technique to classify them according to their disease state.
And then we need to be able to take all that data and convince the FDA in the context of the clinical trial, does this diagnostic actually work for predicting the disease? That really highlights the gap between what we do in research and what could be used directly for clinical use, because the headlines about that paper were very exciting in telling you that the microbiome is more predictive than calprotectin for IBD status. And that’s absolutely true in the context of our research study.
But that doesn’t mean that you can take one sample from your patient and then get their microbiome done by whatever method you choose and be able to do that same readout, anymore than you’ll be able to say compare CBC results that you did with a whole lot of different methods and then try to integrate them into one tracking series over time. There’s so much equal variability that it doesn’t work.
DrMR: Yeah, you got to be cautious of headlines. I mean, headlines are headlines because they’re meant to be catchy and turn your eye. But you have to be tempered and careful, because, man, if you make clinical decisions based upon headlines, you’re going to be pretty off the market, in my opinion. So thank you for making that clarification.
Microbial Insufficiencies Cannot Be Diagnosed with GI Mapping
One other thing I want to ask you before it slips my mind, this is a basic question, but I guess it’s questioned a lot, looking at a microbiota mapping result and seeing that someone has potentially lower levels of Bifidobacterium infantis or Lactobacillus acidophilus, as I understand it, that does not predict what someone needs in terms of a probiotic or if they’ll have a better response. Let’s say they’re low in strain X, by giving strain X, they’ll improve. As far as I know, there’s no data supporting that. Do you have any thoughts there?
RK: No. I’m not aware of any data supporting the idea of other microbial insufficiencies that you can diagnose that way and then replace them with an off-the-shelf probiotic. It’s kind of like saying, I went over to my friend’s house and they had a really nice garden and attributing that to the fact that—like they had a bunch of palm trees, so I’m going to try growing some palm trees in my garden and then that will make it nicer.
And then you might have some follow-up questions like, “Doesn’t your friend live in California or do you live in Wisconsin? Isn’t that going to be an issue?” The environments they got probably has this huge impact on what’s going to grow there. So your hypothesis that you’re testing when you take the approach you’re describing as what’s missing is the seed, but what might be missing is the fertilizer or the soil, right? And so you need all of those components to get that microbe to establish in that ecosystem as well as interactions with a whole of other microbes.
DrMR: Yeah. I agree with that 1000%. And one of the things I often comment regarding the microbiota, it’s really a holistic approach in my opinion to improve one’s microbiota. And this maybe why we see, for example, cardiorespiratory fitness predicts microbiota diversity. And one study even showed starting an exercise program can actually shift your microbiota into what we think is a healthier and more diverse direction. So to think that by simply putting in the seeds, as you said, is going to have this dramatic effect is a bit narrow-minded in terms of scope of approach.
RK: Yeah, absolutely. There’s been some very, very interesting research just lately looking at the impact and mindfulness on the microbiome, so certainly taking a whole systems approach is definitely the way to go there.
DrMR: Definitely, definitely. Well, Rob, I know you got to jet here in just a second. Last question for you, is there anything that people can do to help move the science forward? If you were to ask people to do one thing to maybe help with the work that you’re doing, is there one or two things they can do to help move the science forward?
RK: Absolutely. So, there are two things. One is American Gut, we’ve seen a whole lot of kinds of microbiomes out there in the wild. But there’s a lot of groups we’re not reaching, because each path doesn’t support the cost of adding itself to the project, donations that allow us to target underrepresented groups. So for example, more ethnically diverse groups, those are extremely useful for expanding our knowledge base in seeing where the relationships hold among people with disposable income also hold for the rest of the population.
The second thing is we’re starting a big capital campaign at UCSD at the moment. And the microbiome part of that is essentially one thing that’s really useful is to build out those reference data sets. So if you’re interested in a particular disease state and you’re able to sponsor getting together a cohort who will undergo an intervention anyway and then sponsoring the research on that cohort to figure out, does the intervention work and can you predict who it’s going to work for an who it’s not going to work for based on a standardized microbiome protocol beforehand or tracking during procedure, that’s also extremely helpful to building up that research knowledge base that one day is going turn into a clinical test. And we hope one day soon.
DrMR: Awesome, awesome. All right. Well, I know you’ve got to jet. Thank you so much for taking the time. It’s always great hearing you speaking. And I know it sounds kind of simple, but just I think for the audience hearing from your mouth some of these simple truths about not overreaching with what microbiota mapping can offer clinically will do massive volumes to prevent them from wasting their money and, hopefully, then finding their way to the clinicians and the tests and the treatments that can actually help them. So thank you for your work. Please keep me posted on anything new and exciting. Until we come across each other again in a conference, keep doing the good work that you’re doing.
RK: Very well. Thanks again for having me on and thank you for what you’re doing.
DrMR: Thanks, Rob. Take care. Bye.