1 00:00:06,458 --> 00:00:10,500 Welcome to the Proteomics and Proximity podcast, where your co-hosts Dale Yuzuki, 2 00:00:10,500 --> 00:00:13,916 Cindy Lawley and Sarantis Chlamydas from Olink Proteomics 3 00:00:14,250 --> 00:00:18,458 talk about the intersection of proteomics with genomics for drug target discovery, 4 00:00:18,708 --> 00:00:22,750 the application of proteomics to reveal disease biomarkers and current 5 00:00:22,750 --> 00:00:26,875 trends in using proteomics to unlock biological mechanisms. 6 00:00:27,166 --> 00:00:30,583 Here we have your hosts, Dale, Cindy and us. 7 00:00:30,958 --> 00:00:35,375 Thank you for joining us on the Proteomics and Proximity Podcast. 8 00:00:35,708 --> 00:00:38,583 I'm your host Dale Yuzuki with my co-hosts Cindy Lawley 9 00:00:38,916 --> 00:00:41,458 and my other co-hosts are Sarantis Chlamydas. 10 00:00:42,166 --> 00:00:42,625 Great. 11 00:00:42,625 --> 00:00:47,500 This morning we are talking about Empower Genomics with proteomics. 12 00:00:48,000 --> 00:00:50,625 And Cindy, I'd like to ask you the question. 13 00:00:50,875 --> 00:00:54,708 If I'm involved in genomics, why should I add proteomics? 14 00:00:55,583 --> 00:00:58,208 Well, I'm so happy that you asked me about that, Dale. 15 00:00:58,500 --> 00:01:03,666 So, as you know, I tell this story quite a bit, and so I'm delighted 16 00:01:03,833 --> 00:01:08,875 to do it in the context with both of you, because you add so much to this. 17 00:01:08,875 --> 00:01:12,958 But you know, we've got a big project going with the UK Biobank, 18 00:01:14,541 --> 00:01:15,166 the UK 19 00:01:15,166 --> 00:01:18,250 Biobank, of course, being one of the largest 20 00:01:18,833 --> 00:01:21,833 nationally associated population biobanks 21 00:01:21,833 --> 00:01:25,208 in the world with clinical, genetic 22 00:01:25,541 --> 00:01:29,583 and now on a subset of of over 54,000 samples 23 00:01:30,250 --> 00:01:32,750 of proteomic data. 24 00:01:33,041 --> 00:01:35,666 It's an exciting time, I think, to demonstrate 25 00:01:35,666 --> 00:01:40,708 in large populations the value of layering proteomics onto genomics. 26 00:01:40,708 --> 00:01:44,041 And of course we've we've across many cohorts invest 27 00:01:44,041 --> 00:01:49,041 did a lot in genomics as the costs have have gone down. 28 00:01:49,541 --> 00:01:53,750 Now to back up a little bit, the UK Biobank tell me a little bit about it. 29 00:01:54,625 --> 00:01:55,458 Sure. 30 00:01:55,500 --> 00:02:00,875 So, yeah, it's as a as the name implies, it's based in the UK, 31 00:02:01,250 --> 00:02:05,000 it's affiliated with, you know, the UK has, 32 00:02:05,000 --> 00:02:08,625 you know, one of the largest single payer health care systems in the world. 33 00:02:09,083 --> 00:02:12,208 And having a population based biobank 34 00:02:12,583 --> 00:02:16,250 primarily of northern European descent or ancestry, 35 00:02:16,250 --> 00:02:20,125 but certainly representing Asian as well as as African 36 00:02:20,500 --> 00:02:24,083 and African diaspora descent as well Pakistani descent. 37 00:02:24,083 --> 00:02:26,708 There's quite a nice subset of diversity 38 00:02:26,708 --> 00:02:29,875 in the Biobank, but primarily it's it's northern European. 39 00:02:30,625 --> 00:02:34,208 You know, it was it was started what I think 20 years ago. 40 00:02:34,208 --> 00:02:36,000 I actually should know that off the top of my head. 41 00:02:36,000 --> 00:02:42,166 But it was it was started with the promise of being able to characterize the value 42 00:02:42,583 --> 00:02:46,083 of longitudinal information to health care. 43 00:02:46,666 --> 00:02:51,125 And I think Eric Topol says that it takes about 20 years 44 00:02:51,125 --> 00:02:55,041 on average to move something from discovery to the clinic. 45 00:02:55,875 --> 00:02:59,958 He uses the example of the stethoscope 46 00:03:00,750 --> 00:03:04,166 and moving the stethoscope into the clinic took 17 years. 47 00:03:04,333 --> 00:03:08,000 That seems like a pretty simple mechanism, right? 48 00:03:08,000 --> 00:03:09,708 Listening to your heart. 49 00:03:09,708 --> 00:03:13,333 But yet it took a long time for it to be demonstrated and approved 50 00:03:13,875 --> 00:03:16,041 and to get into the clinic and routine use. 51 00:03:16,583 --> 00:03:21,125 So by longitudinal, then you mean I think they recruit what, half a million 52 00:03:21,125 --> 00:03:25,041 individuals and longitudinal means what they follow them over time. 53 00:03:25,375 --> 00:03:27,625 It means that they're able to call them back. 54 00:03:27,625 --> 00:03:28,375 So they're able 55 00:03:28,375 --> 00:03:32,583 they have medical access to their, you know, clinical data over time. 56 00:03:32,583 --> 00:03:38,208 They understand over time what is, you know, what is 57 00:03:39,833 --> 00:03:42,291 the outcomes within this population. 58 00:03:42,583 --> 00:03:47,166 And I think that's incredibly valuable that consenting has changed over time. 59 00:03:47,166 --> 00:03:50,250 So the ability to actually call back wasn't 60 00:03:50,958 --> 00:03:53,750 initially in many of these biobanks. Right. 61 00:03:53,750 --> 00:03:58,416 And so , I think of FinnGen as one of those that another biobank 62 00:03:58,416 --> 00:04:02,500 that's based in Finland, a population health biobank as well. 63 00:04:03,166 --> 00:04:07,791 They really did sort of lead the way with some of the ability to share data 64 00:04:08,333 --> 00:04:12,541 and protect it at the same time, you know, primarily focused on genetics 65 00:04:12,875 --> 00:04:17,708 and I would say UK Biobank as well has led the way in these abilities 66 00:04:17,708 --> 00:04:22,041 to work with both private and public partnerships. 67 00:04:22,041 --> 00:04:26,125 So being able to work with pharma and in this case 68 00:04:26,125 --> 00:04:29,708 with the proteomic data, this initial set of proteomic data 69 00:04:30,208 --> 00:04:35,250 that was initially started with ten pharma partners. 70 00:04:35,750 --> 00:04:38,416 The model was interestingly 71 00:04:38,416 --> 00:04:41,291 based upon the exome sequencing consortium. 72 00:04:41,916 --> 00:04:46,583 So a group of pharma partners came together in order to get some 73 00:04:47,000 --> 00:04:51,958 to get the participants within the UK Biobank exome sequenced. 74 00:04:52,500 --> 00:04:58,250 There's of course also I think it's 150,000 individuals in the UK Biobank 75 00:04:58,250 --> 00:05:01,875 that are also whole genome sequenced, which is pretty phenomenal, right? 76 00:05:02,166 --> 00:05:04,083 That's a huge number. 77 00:05:04,083 --> 00:05:05,666 So a. Massive management. 78 00:05:05,666 --> 00:05:10,166 And yeah, of all of those dollars and time 79 00:05:10,583 --> 00:05:15,500 and and the potential for building analysis tools with such large datasets. 80 00:05:15,500 --> 00:05:18,416 Right. That's that's, you know, goes without saying. 81 00:05:18,416 --> 00:05:21,791 I remember I think it was 2018, I was at ASHG, 82 00:05:21,791 --> 00:05:24,208 which is the American Society for Human Genetics, 83 00:05:24,750 --> 00:05:29,208 and I was blown away by all of the talks that were referencing the UK 84 00:05:29,250 --> 00:05:32,958 Biobank data, building tools, having discoveries. 85 00:05:32,958 --> 00:05:33,416 Right. 86 00:05:33,500 --> 00:05:37,083 I'm excited to see that same 87 00:05:38,041 --> 00:05:40,666 evolution in 88 00:05:40,666 --> 00:05:43,250 in the discussion 89 00:05:43,250 --> 00:05:47,250 with crowdsourcing these data with proteomics. 90 00:05:48,125 --> 00:05:51,583 So to get back to your original question, Dale, what's the value? 91 00:05:52,208 --> 00:05:56,541 I think Karsten Suhre would say that when you have genetic data 92 00:05:56,666 --> 00:05:59,750 and you have disease, that there's a certain power 93 00:05:59,750 --> 00:06:02,333 you have to detect the relationship between the two. 94 00:06:03,041 --> 00:06:06,750 And in some cases, we have smoking guns like BRCA. Right. 95 00:06:06,750 --> 00:06:11,000 So we're able to to see that there's a lot of penetrance for a variant that 96 00:06:11,291 --> 00:06:15,041 that shows up and has a lot of influence on on predisposition to disease. 97 00:06:15,875 --> 00:06:19,583 But for most diseases, it's death by a thousand cuts, right? 98 00:06:19,583 --> 00:06:24,125 Small amounts of influence. 99 00:06:24,625 --> 00:06:27,208 Cindy, this UK Biobank sounds so interesting. 100 00:06:27,208 --> 00:06:29,541 What can you tell me more about it? 101 00:06:29,541 --> 00:06:30,208 Yeah. So. 102 00:06:30,208 --> 00:06:35,708 So the UK Biobank itself is a longitudinal collection of data 103 00:06:35,750 --> 00:06:38,875 I think it started in the mid 2000. 104 00:06:38,875 --> 00:06:45,083 So around 2006 they targeted an age group, sort of middle age group 105 00:06:45,416 --> 00:06:47,208 and they followed them over time 106 00:06:47,208 --> 00:06:51,083 and so it's over half a million individuals within the UK. 107 00:06:51,166 --> 00:06:56,333 Yeah, quite, quite a undertaking to enroll all those participants. 108 00:06:56,333 --> 00:07:01,041 And what I will never forget is when I attended ASHG 109 00:07:01,666 --> 00:07:06,708 the American Society of Human Genetics in in 2018 110 00:07:07,416 --> 00:07:11,208 the number of talks I remember searching on UK BB 111 00:07:11,375 --> 00:07:15,166 as just in a short an acronym and the number of talks 112 00:07:15,166 --> 00:07:19,583 talking about using the UK B data, particularly genetic data 113 00:07:20,291 --> 00:07:22,708 for validating clinical findings 114 00:07:22,708 --> 00:07:25,791 was there were a lot of talks. 115 00:07:25,791 --> 00:07:29,291 So it's always been high on Illumina's 116 00:07:29,291 --> 00:07:33,083 radar and it's very high on all of those sequencing 117 00:07:33,083 --> 00:07:37,458 technology innovator radar, innovator’s radars 118 00:07:38,500 --> 00:07:41,208 like Thermo Fisher, of course, and all of those library 119 00:07:41,208 --> 00:07:44,750 prep companies that have different methods of library prep. 120 00:07:45,000 --> 00:07:47,666 Have they already sequenced everybody? 121 00:07:47,666 --> 00:07:52,416 So they've got a whole genome sequence on, I think it's about 150,000 individuals 122 00:07:52,416 --> 00:07:56,500 that was primarily led by Decode Genetics. 123 00:07:56,500 --> 00:07:58,583 As I understand it, 124 00:07:58,583 --> 00:08:01,750 the publication is pretty recent actually. 125 00:08:01,791 --> 00:08:03,833 So it's a 126 00:08:04,291 --> 00:08:06,208 there's a lot to dig into. 127 00:08:06,208 --> 00:08:09,666 It's a pretty phenomenal dataset. 128 00:08:10,000 --> 00:08:16,083 The bulk of the sequencing for most samples, I believe is exome sequencing. 129 00:08:16,083 --> 00:08:17,000 That's my understanding. 130 00:08:17,000 --> 00:08:20,541 So I think it's it's over 450,000 individuals. 131 00:08:20,541 --> 00:08:24,125 So I think for some they've got both. 132 00:08:24,208 --> 00:08:26,541 And since it's a single payer system 133 00:08:26,541 --> 00:08:29,625 with the electronic records of NHS, 134 00:08:29,958 --> 00:08:33,708 that means they can drill down into exactly right 135 00:08:33,708 --> 00:08:38,750 their whole exome sequence and whatever condition they may have 136 00:08:38,750 --> 00:08:40,083 And this is an ongoing thing. 137 00:08:40,083 --> 00:08:42,541 Is that right? Cancer, diabetes? 138 00:08:42,833 --> 00:08:43,375 That's right. 139 00:08:43,375 --> 00:08:46,500 And the ability to actually return results to those patients, 140 00:08:46,500 --> 00:08:48,583 I think has evolved over time. Right. 141 00:08:48,583 --> 00:08:49,666 Because that costs money 142 00:08:49,666 --> 00:08:53,750 to set expectations, make sure that we're we're communicating, 143 00:08:55,166 --> 00:08:57,250 you know, in a way that's best practices. 144 00:08:57,250 --> 00:09:00,625 So I think that the UK Biobank has spearheaded 145 00:09:00,625 --> 00:09:03,666 a lot of our understanding about 146 00:09:03,666 --> 00:09:06,666 best practices there as well. 147 00:09:06,833 --> 00:09:08,666 And as far as.... Go ahead. 148 00:09:08,666 --> 00:09:11,166 Oh, I was just going to say so back to your original question 149 00:09:11,166 --> 00:09:14,333 about what's the value of layering proteomics onto genomics. 150 00:09:14,750 --> 00:09:20,000 This is a great example where an enormous amount of investment has gone 151 00:09:20,000 --> 00:09:24,083 into collecting genetic information for this very valuable population, 152 00:09:24,708 --> 00:09:30,541 with advances in diagnostics, in guiding cancer treatment. 153 00:09:30,541 --> 00:09:33,958 So a lot of these advances have made it to the clinic already, 154 00:09:34,000 --> 00:09:36,041 which is pretty phenomenal. 155 00:09:36,041 --> 00:09:39,416 And that's been driven, you know, globally, it's it's exciting 156 00:09:39,791 --> 00:09:42,416 where proteomics fits in or the way that I think of it. 157 00:09:42,416 --> 00:09:43,291 I was 158 00:09:44,083 --> 00:09:48,958 I had a conversation with Karsten Suhre, who's at Weill Cornell in Qatar 159 00:09:48,958 --> 00:09:54,041 and in New York and he he really I had an aha moment with him. 160 00:09:54,041 --> 00:09:58,916 He he essentially would say that an intermediate phenotype like proteomics 161 00:09:59,625 --> 00:10:03,625 acts to magnify the effect between genetics and disease. 162 00:10:03,625 --> 00:10:04,166 So, of course, 163 00:10:04,166 --> 00:10:07,916 we've been looking for these associations between genetics and disease 164 00:10:09,000 --> 00:10:10,958 since we've been collecting genetic information. 165 00:10:10,958 --> 00:10:13,041 And some of those 166 00:10:14,125 --> 00:10:16,166 links are hard to see because 167 00:10:16,166 --> 00:10:19,375 we need so many samples to be able to see them. 168 00:10:19,791 --> 00:10:23,416 And so as we've increased the numbers of samples like in the UK Biobank, 169 00:10:23,666 --> 00:10:27,208 we're able to make these associations more clearly. 170 00:10:27,958 --> 00:10:32,583 And I'll say, you know, we wished early on we hoped for smoking 171 00:10:32,583 --> 00:10:36,291 guns for a lot of diseases and we we did see a few of them. 172 00:10:36,291 --> 00:10:36,625 Right. 173 00:10:36,625 --> 00:10:41,833 So there's certainly PCSK9 for familial hypercholesterolemia. 174 00:10:41,833 --> 00:10:44,750 There are some standard ones, BRCA for breast cancer. 175 00:10:45,625 --> 00:10:50,666 There's there's some examples where we have a lot of penetrance or a lot of, 176 00:10:50,666 --> 00:10:55,208 you know, a lot of affect on someone's likelihood of getting a disease 177 00:10:55,875 --> 00:11:01,083 from single variants or single loci or single genes. 178 00:11:01,083 --> 00:11:04,708 But for most diseases like Type 2 diabetes, cardiovascular disease 179 00:11:04,708 --> 00:11:08,458 this is a death by a thousand cuts, meaning lots of variants. 180 00:11:08,458 --> 00:11:12,291 Give a little tiny effect in changing our risk. 181 00:11:12,291 --> 00:11:16,083 And so that's where that's where having having the ability 182 00:11:16,083 --> 00:11:19,250 to amplify or to put a magnifying glass on those 183 00:11:19,625 --> 00:11:24,208 those relationships between genetics and disease is incredibly useful. 184 00:11:24,208 --> 00:11:25,333 So I think, you know, a 185 00:11:25,333 --> 00:11:29,375 ton of of work has been done in proteomics and cardiovascular disease. 186 00:11:29,375 --> 00:11:32,250 And I think many advances have happened there. 187 00:11:32,250 --> 00:11:35,541 Now getting back to the UK Biobank. 188 00:11:35,541 --> 00:11:38,833 You mentioned before that recently they were working with Olink 189 00:11:39,000 --> 00:11:42,916 then to look at the proteome of tens of thousands of individuals. 190 00:11:42,916 --> 00:11:43,916 Yeah. Yeah. 191 00:11:43,916 --> 00:11:47,875 I would say it's not just the UK Biobank, but but 13 pharma partners. 192 00:11:47,875 --> 00:11:48,166 Right. 193 00:11:48,166 --> 00:11:53,875 So, so it certainly required consent and partnership with the UK Biobank, 194 00:11:54,208 --> 00:11:55,875 just like the exome sequencing 195 00:11:55,875 --> 00:11:59,458 consortium was was done in collaboration with the UK Biobank. 196 00:11:59,833 --> 00:12:03,625 But the access to the technology, just like with the exome 197 00:12:03,625 --> 00:12:07,541 sequencing consortium, access to the technology was spearheaded 198 00:12:07,541 --> 00:12:10,250 by pharma partners that were very keen 199 00:12:10,791 --> 00:12:15,750 to build a structure for a more, I like to say, a systematic approach 200 00:12:16,083 --> 00:12:19,791 to therapeutic target discovery, not only biomarker discovery, 201 00:12:19,791 --> 00:12:23,541 which is sort of traditional proteomics, but to to therapeutic target discovery, 202 00:12:23,541 --> 00:12:28,291 which I think is enabled by genetics, proteomics as well as clinical data. 203 00:12:29,000 --> 00:12:31,750 So we're getting back to this idea of empowering genomics 204 00:12:31,750 --> 00:12:33,541 with proteomics, right? 205 00:12:33,541 --> 00:12:35,625 What can you tell me about that? 206 00:12:35,625 --> 00:12:35,958 Yeah. 207 00:12:35,958 --> 00:12:37,166 So I think there's this, 208 00:12:37,166 --> 00:12:41,083 you know, I immediately think of that Karsten Suhre, magnifying glass, right. 209 00:12:41,791 --> 00:12:44,291 But the, the UK Biobank’s 210 00:12:44,333 --> 00:12:48,125 initial findings which are in a preprint that came out 211 00:12:48,125 --> 00:12:52,041 in June, middle of June, that’s on bioRxiv. 212 00:12:52,375 --> 00:12:56,250 Their initial paper really just was scratching 213 00:12:56,250 --> 00:12:59,875 the surface of what's possible with this enormous dataset. 214 00:12:59,875 --> 00:13:04,000 So their first paper was about 1500 proteins. 215 00:13:04,000 --> 00:13:08,000 So our first, our first product that Olink first product 216 00:13:08,000 --> 00:13:10,791 on the Explore platform that has the NGS readout. 217 00:13:11,375 --> 00:13:16,375 So they use that first tranche of proteins across 54,000 samples 218 00:13:16,916 --> 00:13:19,583 and the first really the bulk of 219 00:13:19,583 --> 00:13:24,416 those data are to look at correlations between gene regions, 220 00:13:24,958 --> 00:13:27,208 you know, and the genotypes in those gene regions. 221 00:13:27,541 --> 00:13:28,666 And protein levels. 222 00:13:28,666 --> 00:13:30,750 So really just looking what are the correlations? 223 00:13:30,750 --> 00:13:34,708 What's the list of all the possible relationships between genetic regions 224 00:13:35,291 --> 00:13:38,458 and protein levels that might be elucidated 225 00:13:38,666 --> 00:13:41,916 and examined further in this beautiful dataset? 226 00:13:42,916 --> 00:13:43,541 What I 227 00:13:43,541 --> 00:13:46,791 you know, and I'll speculate on what I think they're going to be doing next 228 00:13:46,791 --> 00:13:51,291 and what my guess is that they're very deep into doing this within 229 00:13:52,041 --> 00:13:56,125 these companies is to then do Mendelian randomization, 230 00:13:56,125 --> 00:13:58,375 which is a statistical approach to kind of 231 00:13:58,916 --> 00:14:02,458 determine which of these relationships, which of these correlations 232 00:14:02,458 --> 00:14:05,875 between gene regions and and protein levels. 233 00:14:06,416 --> 00:14:10,250 You know, when you put it when you bring in the clinical data on disease, 234 00:14:10,625 --> 00:14:15,500 which of these hold up as being unlikely to be happening by chance alone? 235 00:14:15,958 --> 00:14:21,500 So now you sort of have the the ones that are likely just, you know, coincidence. 236 00:14:21,500 --> 00:14:24,833 I mean, they might still be important, but but let's just I think pharma 237 00:14:24,833 --> 00:14:29,750 would like to have ten great targets, over 400 238 00:14:30,750 --> 00:14:31,375 unsure 239 00:14:31,375 --> 00:14:34,375 targets, because that that's a lot of rabbit holes to go down. 240 00:14:34,375 --> 00:14:37,041 So if they can narrow it down, then 241 00:14:37,500 --> 00:14:42,041 then I think there's a lot of excitement around being able 242 00:14:42,041 --> 00:14:46,958 to have some quick wins with proteomics, genomics and clinical data. 243 00:14:47,208 --> 00:14:49,208 Well, to back up just a little bit, 244 00:14:50,000 --> 00:14:53,125 you're talking about Mendelian randomization, you're talking 245 00:14:53,125 --> 00:14:59,125 about genomic data in terms of a whole exome of 10,000 people or 50,000 people, 246 00:14:59,791 --> 00:15:02,625 and now you're talking about 1500 proteins. 247 00:15:02,875 --> 00:15:04,875 Can you walk me through that a little bit? 248 00:15:04,875 --> 00:15:05,583 Yeah, sure. 249 00:15:05,583 --> 00:15:10,166 So so when you're looking at the genetic data and here we've got, 250 00:15:11,416 --> 00:15:14,458 you know, some whole genome sequencing data as well as exome sequencing data. 251 00:15:14,458 --> 00:15:18,750 So you can imagine you have a list of ways or places in the genome almost like 252 00:15:19,583 --> 00:15:24,458 like geographic locations, almost GPS coordinates in the 253 00:15:24,500 --> 00:15:28,833 on the chromosomes where we know they vary across the samples. 254 00:15:29,125 --> 00:15:32,791 So those variable regions, we’ll call them SNPs, you know, 255 00:15:33,000 --> 00:15:37,000 that that's the term that that we use for the simplest kind of variation, 256 00:15:37,000 --> 00:15:39,750 just single base pair variation, but we'll just call them snips 257 00:15:40,041 --> 00:15:43,458 because, because there can be other kinds of variation that are captured there too. 258 00:15:43,791 --> 00:15:48,208 But if you look at the variance, it's just a single variant within the genome. 259 00:15:48,666 --> 00:15:50,125 You can look at the 260 00:15:51,166 --> 00:15:53,958 the representation of what people's genotype 261 00:15:53,958 --> 00:15:58,625 is at that location, and you can look at every single protein 262 00:15:58,625 --> 00:16:01,916 in that 1500 protein list and see, 263 00:16:01,916 --> 00:16:06,208 do we have a significant correlation between the genotype 264 00:16:06,791 --> 00:16:08,000 and the protein level? 265 00:16:08,000 --> 00:16:09,500 So that's sort of the first step. 266 00:16:09,500 --> 00:16:12,250 That's a lot of tests write comparisons. 267 00:16:12,833 --> 00:16:16,500 Cindy, so I understand these SNPs can also be outside of the gene. 268 00:16:16,500 --> 00:16:17,750 Right would be also make it 269 00:16:17,750 --> 00:16:21,375 So they could be regulatory regions. Absolutely. It’s not necessarily falling into the gene 270 00:16:21,625 --> 00:16:25,250 Is there any threshold to what they are checking, what region they check in the gene? 271 00:16:25,625 --> 00:16:26,041 Yeah. 272 00:16:26,041 --> 00:16:31,583 So there's both a statistical threshold that they accept as a as a standard. 273 00:16:31,583 --> 00:16:35,833 But also when you're doing so many tests, you have to correct 274 00:16:35,833 --> 00:16:37,083 for multiple tests, right? 275 00:16:37,083 --> 00:16:40,375 Because the more tests you do, the you're increasing your chances 276 00:16:40,375 --> 00:16:43,083 of seeing a false positive. So adjusting for that 277 00:16:44,333 --> 00:16:45,750 is something that, you know, 278 00:16:45,750 --> 00:16:50,166 we go through peer review to make sure we have best practices and agreement on. 279 00:16:50,375 --> 00:16:53,541 I mean, these statistical associations are massive. 280 00:16:53,541 --> 00:16:57,666 I mean, in a given single individual's whole genome, you're looking at maybe 281 00:16:57,666 --> 00:16:59,916 four million SNPs? Yeah, right. 282 00:17:00,291 --> 00:17:02,958 You have 4 million SNPs and then you’ve got 283 00:17:03,333 --> 00:17:07,375 1500 proteins you're associating those with, if I understand you correctly. 284 00:17:07,375 --> 00:17:07,791 Yeah. 285 00:17:07,791 --> 00:17:12,583 And then you multiply this times what they did, 54,000 individuals. 286 00:17:12,791 --> 00:17:15,333 They did 54,000. So I mean. 287 00:17:15,625 --> 00:17:18,625 And they discovered, you know, about 10,000, 288 00:17:18,625 --> 00:17:24,166 I think it was around 10,200 relationships between gene regions and protein levels. 289 00:17:24,166 --> 00:17:26,083 Right. That's a massive number. 290 00:17:26,083 --> 00:17:29,541 So that's those those could many of those could just be coincidence, right? 291 00:17:29,541 --> 00:17:31,625 Just correlations, not causation. Right. 292 00:17:31,625 --> 00:17:35,208 We were all familiar with that, that phrase. 293 00:17:35,541 --> 00:17:40,541 So so 85% of those relationships were novel. 294 00:17:41,916 --> 00:17:43,708 Now, the relationship. 295 00:17:43,708 --> 00:17:46,666 Cindy, could they be both in cis and trans, both of them? 296 00:17:46,666 --> 00:17:48,291 Correlation or. 297 00:17:49,208 --> 00:17:49,916 What's that? 298 00:17:49,916 --> 00:17:53,666 Sarantis. Cis and trans would be like this correlations could be there. 299 00:17:53,833 --> 00:17:54,833 So yeah. 300 00:17:54,833 --> 00:17:59,125 So these correlations can be what we call cis or they can be in trans. 301 00:17:59,666 --> 00:18:02,583 So cis just is getting back to Dale’s question 302 00:18:02,583 --> 00:18:06,666 about whether or actually was your question Sarantis, about whether 303 00:18:06,666 --> 00:18:10,750 these variants, these SNPs are inside genes 304 00:18:11,375 --> 00:18:13,500 or are they outside genes? 305 00:18:14,375 --> 00:18:17,458 And if they're in genes or or in close proximity 306 00:18:17,458 --> 00:18:20,458 to the genes, that code for the protein itself. 307 00:18:21,041 --> 00:18:21,416 Right. 308 00:18:21,416 --> 00:18:26,958 So you've got a variant that's in a gene coding for a protein. 309 00:18:27,375 --> 00:18:30,416 If you see a correlation that's significant between 310 00:18:30,416 --> 00:18:34,375 those two, we call that a ”cis-pQTL” and that’s a feel 311 00:18:34,375 --> 00:18:38,250 good measure that says, oh, we must be measuring the right protein. 312 00:18:38,250 --> 00:18:42,250 If this is real, then and there's ways to to press on it, 313 00:18:42,250 --> 00:18:46,500 to check it and validate it, of course, with us orthogonal data, but that's it. 314 00:18:46,791 --> 00:18:52,166 So people often talk about cis-pQTL discoveries being verification 315 00:18:52,166 --> 00:18:54,791 then of of having measured the right protein 316 00:18:55,125 --> 00:18:59,541 because of course our assay is not a Mass Spec method. 317 00:18:59,541 --> 00:19:02,083 We're using antibodies as hooks. 318 00:19:02,083 --> 00:19:05,708 We're using two antibodies as a hook to hook 319 00:19:05,708 --> 00:19:09,208 a protein out of out of solution. 320 00:19:09,958 --> 00:19:13,375 And we have little single stranded oligos attached to them. 321 00:19:13,375 --> 00:19:18,458 So those oligos can then hybridize, we can extend and amplify 322 00:19:18,458 --> 00:19:22,500 that up just like any old library prep for for sequencing. 323 00:19:22,500 --> 00:19:24,333 And then we count those oligos 324 00:19:24,333 --> 00:19:27,958 as a proxy for the original level of the proteins in the sample. 325 00:19:28,500 --> 00:19:32,208 And so when you're doing an affinity method, right, a hooking method 326 00:19:32,250 --> 00:19:36,083 to pull it out, not only is it great for low abundant proteins, 327 00:19:36,083 --> 00:19:39,416 that's one of the things we add value to with with mass spec folks. 328 00:19:39,416 --> 00:19:42,583 They, they like us because they can look at areas of the proteome like 329 00:19:42,583 --> 00:19:46,458 they couldn't see easily with mass spec, without tons of sample 330 00:19:46,458 --> 00:19:49,375 and a lot of a lot of control of variability. 331 00:19:49,916 --> 00:19:53,750 So it's, it's, it's a nice method from that perspective, 332 00:19:54,166 --> 00:19:57,125 but it's, it's a little bit indirect because we're, 333 00:19:57,125 --> 00:20:00,291 we're pulling out the protein and converting it to DNA signal. 334 00:20:00,625 --> 00:20:03,500 So making sure we have a way to normalize those data 335 00:20:04,041 --> 00:20:07,916 and an end, you know, just like with mass spec in any proteomics experiment 336 00:20:07,916 --> 00:20:10,208 to manage variability 337 00:20:11,125 --> 00:20:15,291 from batch to batch, you know, these are important aspects that that 338 00:20:15,291 --> 00:20:20,000 proteomics scientists are much better prepared to describe or explain than I am. 339 00:20:20,250 --> 00:20:23,000 I have come to appreciate it. 340 00:20:23,000 --> 00:20:25,541 Now, Cindy, something that you touched upon, 341 00:20:25,583 --> 00:20:29,666 right, was the sort of drug discovery dimension of this. 342 00:20:30,000 --> 00:20:33,916 But even before we go in that direction, you also mentioned something 343 00:20:33,916 --> 00:20:36,291 in terms of SNPs and genes, 344 00:20:37,208 --> 00:20:40,041 the majority of GWAS, thousands, right. 345 00:20:40,041 --> 00:20:42,291 3500 GWAS studies or 346 00:20:43,250 --> 00:20:45,208 many, many, many. 347 00:20:45,208 --> 00:20:48,500 Oftentimes these SNPs that are associated with risk. 348 00:20:48,625 --> 00:20:51,958 Often our gene deserts are there in. Right. 349 00:20:51,958 --> 00:20:53,583 There's no function. 350 00:20:53,583 --> 00:20:54,416 That's right. 351 00:20:54,416 --> 00:20:56,000 What can you comment on that? Right. 352 00:20:56,000 --> 00:20:57,958 So these pQTLs. Right. 353 00:20:57,958 --> 00:21:03,625 are SNPs, but aren’t they just random, so to speak random places in the genome? 354 00:21:04,041 --> 00:21:05,333 Yeah, so good question. 355 00:21:05,333 --> 00:21:10,791 So I'm going to reference a paper by Lasse Folkersen and Anders Mälarstig. 356 00:21:11,291 --> 00:21:13,791 Now the two of them, along with collaborators, 357 00:21:13,791 --> 00:21:16,958 there's a long list of authors that I won't I won't list. 358 00:21:17,291 --> 00:21:20,750 Brilliant, obviously across multiple cohorts, 359 00:21:21,166 --> 00:21:25,041 they have their milestone paper within a study 360 00:21:25,041 --> 00:21:28,083 called the Scallop Study, which is really a cohort of cohorts. 361 00:21:29,541 --> 00:21:33,916 They were doing what the UK Biobank wants to do. 362 00:21:34,250 --> 00:21:36,791 This is me putting words in the UK Biobank’s mouth. 363 00:21:36,833 --> 00:21:41,041 But I think that the Folkersen et al. milestone 364 00:21:41,041 --> 00:21:45,500 publication, is a powerful precursor to what the UK Biobank 365 00:21:46,500 --> 00:21:48,083 is, 366 00:21:48,083 --> 00:21:52,000 has the possibility to do. In Folkersen et al. 367 00:21:52,041 --> 00:21:56,250 they looked at just 90 proteins I say just although at the time in 2020 368 00:21:56,250 --> 00:21:58,708 that was a lot of multiplex proteins of course 369 00:21:59,291 --> 00:22:03,291 they looked at cardio, primarily cardiovascular, what we what we broadly 370 00:22:03,291 --> 00:22:08,166 categorize as cardiovascular proteins and they did the same kind of study. 371 00:22:08,166 --> 00:22:11,875 So on 30,000 samples, they looked at 90 proteins 372 00:22:12,041 --> 00:22:15,291 with genetic, clinical and proteomic data. 373 00:22:16,333 --> 00:22:20,166 They did the correlations just like the UK Biobank has done in their preprint. 374 00:22:20,875 --> 00:22:24,208 The 90 proteins resulted in 450. 375 00:22:24,208 --> 00:22:27,041 Yea, a little over 450 pQTLs 376 00:22:27,791 --> 00:22:31,375 Some of those are cis-pQTLs as Sarantis hints on 377 00:22:31,375 --> 00:22:31,708 Right. 378 00:22:31,708 --> 00:22:36,166 So 88% of the proteins had cis-pQTLs identified there 379 00:22:36,416 --> 00:22:39,083 That's like I said, a feel good method, a method 380 00:22:39,958 --> 00:22:44,375 or something we can kind of point to, to say this, this, this looks like it's, 381 00:22:45,000 --> 00:22:46,500 you know, increasing our confidence 382 00:22:46,500 --> 00:22:50,125 that we're measuring the right protein, although there are good biological reasons 383 00:22:50,125 --> 00:22:54,833 why you might not see a cis-pQTL, but the remaining 384 00:22:56,208 --> 00:22:58,666 trans-pQTLs were 385 00:22:59,458 --> 00:23:04,375 essential discovery of trans-pQTLs is incredibly important 386 00:23:04,708 --> 00:23:06,833 to understand protein-protein interactions. 387 00:23:06,833 --> 00:23:11,625 So I may have taken a bit of a meandering way to get back to your question, Dale, 388 00:23:11,625 --> 00:23:15,458 about these relationships but trans-pQTLs 389 00:23:16,541 --> 00:23:18,500 and figuring out where 390 00:23:18,500 --> 00:23:22,375 those are coding, you know, what proteins 391 00:23:22,375 --> 00:23:26,041 are they coding for, what gene regions are they associated with? 392 00:23:26,125 --> 00:23:29,125 It's not a trivial matter. 393 00:23:29,125 --> 00:23:32,541 And so I've had discussions with Lasse as well as Anderson 394 00:23:32,916 --> 00:23:35,666 or sorry as well as Anders around this 395 00:23:36,958 --> 00:23:38,000 this challenge. 396 00:23:38,000 --> 00:23:42,208 And so just to define trans-pQTLs… so as a reminder cis-pQTLs 397 00:23:42,208 --> 00:23:47,291 are where you have the gene variant is either in or near 398 00:23:47,666 --> 00:23:52,041 the gene that codes for the protein that you're measuring. 399 00:23:52,041 --> 00:23:55,375 So the relationship between them, the correlation is between 400 00:23:55,375 --> 00:23:59,958 the gene region and the protein itself that’s a cis-pQTL 401 00:23:59,958 --> 00:24:03,083 So you see you have say you have a particular protein 402 00:24:03,083 --> 00:24:04,750 your measuring will say 403 00:24:05,708 --> 00:24:09,916 TNF-alpha, or alpha-TNF and there’s a SNP 404 00:24:10,166 --> 00:24:12,625 Then that codes for alpha TNF 405 00:24:13,000 --> 00:24:16,791 is in the same chromosome within, I don't know. 406 00:24:17,125 --> 00:24:19,666 A couple hundred pairs of a. Million base pairs. 407 00:24:19,666 --> 00:24:20,333 So yeah. 408 00:24:20,333 --> 00:24:20,750 When you're. 409 00:24:20,750 --> 00:24:25,583 In the general in the general region and so there could be 410 00:24:25,583 --> 00:24:30,291 those million base pairs a lot of other genes, but nonetheless. 411 00:24:30,333 --> 00:24:30,708 Right. 412 00:24:30,708 --> 00:24:34,250 You're saying that that particular snip was controlling alpha TNF. 413 00:24:34,458 --> 00:24:36,916 It's suggests that I think they might not. 414 00:24:36,958 --> 00:24:41,166 Yeah, they may not say it quite so strongly simply because there's you. 415 00:24:41,166 --> 00:24:42,791 Know it's association. 416 00:24:42,791 --> 00:24:44,208 Yeah. Yeah exactly. It's just. 417 00:24:44,208 --> 00:24:45,833 statistical calculation. Got it. 418 00:24:45,833 --> 00:24:49,583 and so with a trans-pQTL, what that is, is you’ve got a variant 419 00:24:49,875 --> 00:24:52,833 you know, you might have a gene coding for a protein 420 00:24:53,666 --> 00:24:55,750 and that gene might be on chromosome nine. 421 00:24:56,416 --> 00:25:00,833 But you might have the pQTL on chromosome 19. 422 00:25:01,291 --> 00:25:02,000 You know, you might have it 423 00:25:02,000 --> 00:25:05,708 on a completely different chromosome, a correlation with that same protein. 424 00:25:06,291 --> 00:25:10,041 So the sort of Occam's Razor, you know, the easiest 425 00:25:10,041 --> 00:25:11,958 the most straightforward possibility 426 00:25:11,958 --> 00:25:15,291 is that that there's a relationship between those two proteins, right? 427 00:25:15,291 --> 00:25:17,500 That there’s protein-protein interaction going on there. 428 00:25:17,500 --> 00:25:21,708 And in fact, the STRING database is a publicly available database 429 00:25:21,750 --> 00:25:27,208 that records and collects and is curated around protein-protein interactions. 430 00:25:27,208 --> 00:25:30,833 And so what the team would do, you know, in asking them how they 431 00:25:31,708 --> 00:25:34,250 how do they dig into each of these relationships? 432 00:25:34,791 --> 00:25:38,458 And what they would do is look, they report the closest 433 00:25:38,458 --> 00:25:42,208 gene to the location that's in trans with this protein. 434 00:25:42,833 --> 00:25:46,125 They they report the closest gene geographically. 435 00:25:46,125 --> 00:25:47,666 And then they also report 436 00:25:47,666 --> 00:25:51,625 because they do kind of a deep dive into surrounding genes, as you say Dale. 437 00:25:51,625 --> 00:25:55,625 There could be, you know, surrounding genes that that might be implicated. 438 00:25:55,625 --> 00:25:58,916 They look at those other surrounding genes and they say, you know, 439 00:25:58,916 --> 00:26:02,666 what's the shortest pathway back to that protein? 440 00:26:02,916 --> 00:26:07,791 And that is a fascinating conversation because once you once you put together 441 00:26:07,791 --> 00:26:12,541 a pathway analysis like that and we talk about different diseases, 442 00:26:12,875 --> 00:26:17,083 now you've got some pathways in, say, Alzheimer's disease 443 00:26:17,958 --> 00:26:20,208 and you’ve got some pathways in, say, schizophrenia. 444 00:26:20,208 --> 00:26:22,250 I'm just picking two neurological diseases. 445 00:26:22,625 --> 00:26:27,416 And now if you can imagine a Venn diagram of the pathways those two have in common, 446 00:26:28,416 --> 00:26:29,666 and that is 447 00:26:29,666 --> 00:26:33,666 an opportunity for us to understand the mechanistic biology 448 00:26:34,000 --> 00:26:37,916 that's in common between those two neurological diseases, if any. 449 00:26:37,916 --> 00:26:41,000 You know, I'm just picking those out of the air. 450 00:26:41,166 --> 00:26:45,083 If we can return back to that Folkersen landmark paper. 451 00:26:45,291 --> 00:26:46,166 Mm hmm. 452 00:26:46,708 --> 00:26:49,416 So, if I understand correctly, there were 90 proteins. 453 00:26:49,708 --> 00:26:52,500 How many tens of thousands of samples? 454 00:26:52,500 --> 00:26:53,750 30,000 samples. Just over. 455 00:26:53,750 --> 00:26:56,916 Okay, so 30,000 samples times 90 proteins. 456 00:26:56,916 --> 00:27:00,958 And they also had like whole genome data on those 30,000 individuals. 457 00:27:00,958 --> 00:27:03,000 Is that. Right? They had genetic data that you could. 458 00:27:03,000 --> 00:27:04,791 So I don't know that. 459 00:27:04,791 --> 00:27:07,250 Remember, this is a cohort of cohorts. 460 00:27:07,250 --> 00:27:10,875 So I think they had GWAS data or genotyping data, 461 00:27:10,875 --> 00:27:15,125 you know, array data on some of those and sequencing data on others. 462 00:27:15,125 --> 00:27:18,500 I wouldn't want to represent that, but my guess is that they 463 00:27:19,000 --> 00:27:22,041 they had variation, genetic data that they had in common. 464 00:27:22,041 --> 00:27:22,333 Right. 465 00:27:22,333 --> 00:27:23,291 Because you can convert 466 00:27:23,291 --> 00:27:27,166 a whole genome sequencing dataset to a list of variants 467 00:27:27,166 --> 00:27:28,958 Understood. And yeah. Right. 468 00:27:28,958 --> 00:27:32,541 So they had all the genetic data of 30,000 individuals. 469 00:27:32,541 --> 00:27:34,541 They looked at these 90 proteins 470 00:27:35,000 --> 00:27:38,458 and then you mentioned that they're able to connect it then to disease. 471 00:27:38,666 --> 00:27:41,250 Yeah. So so you do the same thing. 472 00:27:41,250 --> 00:27:44,916 This looked at relationships between genetic, 473 00:27:45,333 --> 00:27:47,375 you know, state and protein levels. 474 00:27:47,375 --> 00:27:49,708 So you look for all those correlations. 475 00:27:49,708 --> 00:27:54,166 In this paper, they found 450 pQTLs that exceeded 476 00:27:54,166 --> 00:27:55,916 their significance threshold. 477 00:27:55,916 --> 00:27:59,375 And you could, you know, as you touched on before there, you know, 478 00:27:59,375 --> 00:28:01,750 that's why we have peer review to make sure that we're not 479 00:28:02,625 --> 00:28:06,208 that we're held accountable for the number of tests that we're doing, 480 00:28:06,208 --> 00:28:08,375 that we're you know, we're really trying to be 481 00:28:08,375 --> 00:28:12,458 as as transparent as possible and publishing these data. 482 00:28:12,458 --> 00:28:16,041 And by the way, it was published in Nature Metabolism in 2020. 483 00:28:16,666 --> 00:28:17,666 So once you 484 00:28:17,666 --> 00:28:21,416 you see all the correlations, imagine you have this list of correlations. 485 00:28:21,958 --> 00:28:24,958 You can layer those clinical data then in. 486 00:28:24,958 --> 00:28:27,416 So now you know the disease information 487 00:28:28,000 --> 00:28:30,833 and you can look at these different sets of data. 488 00:28:30,833 --> 00:28:33,750 So genetics, proteomics and disease 489 00:28:34,333 --> 00:28:37,916 and you can sample from these and determine how often 490 00:28:37,916 --> 00:28:40,833 with the relationships between three of these units, 491 00:28:41,250 --> 00:28:43,916 how often would that happen by chance alone? 492 00:28:43,916 --> 00:28:45,333 If it would happen by chance alone? 493 00:28:45,333 --> 00:28:46,250 Quite often. 494 00:28:46,250 --> 00:28:48,500 Then we let that fall away. 495 00:28:49,166 --> 00:28:52,708 If it seems quite unusual to see these relationships, 496 00:28:52,708 --> 00:28:56,750 then those are the ones that we elevate to potential causality. 497 00:28:57,125 --> 00:29:01,583 And so in this paper they elevated from the 450 relationships correlations. 498 00:29:01,583 --> 00:29:06,208 They elevated 25 that they suggest appear causal. 499 00:29:06,791 --> 00:29:10,916 And some of those examples I think, are validated. 500 00:29:10,916 --> 00:29:17,500 All I know are validated clinical targets for existing therapies, super exciting 501 00:29:17,500 --> 00:29:20,500 because then it's like, oh, looks like we're on the right track, right? 502 00:29:21,000 --> 00:29:22,583 And then of course some novel findings. 503 00:29:22,583 --> 00:29:27,000 So they they report 14 validated clinical targets, 504 00:29:27,000 --> 00:29:30,750 known clinical targets like CASP-8 in breast cancer was one of them that 505 00:29:31,791 --> 00:29:32,833 I can think of. 506 00:29:32,833 --> 00:29:36,625 So CASP-8 is something known already before to be involved in breast cancer? 507 00:29:36,625 --> 00:29:39,458 That’s right. And then they rediscovered it? 508 00:29:39,458 --> 00:29:43,083 Yea, CASP-8 is a known therapeutic target in breast cancer. 509 00:29:43,083 --> 00:29:44,041 I see. 510 00:29:44,041 --> 00:29:48,416 And then 11 of those were novel, so they were not able to see any evidence 511 00:29:48,416 --> 00:29:53,541 of 11 of their findings that elevated again to causality, potential causality. 512 00:29:54,000 --> 00:29:57,750 And those are the exciting ones for for a new programs potentially 513 00:29:58,375 --> 00:30:00,833 and then and then 18 514 00:30:00,875 --> 00:30:04,666 they they reported 18 potential repurposing opportunities. 515 00:30:04,666 --> 00:30:08,750 So that's super exciting to me because if you've got an existing drug 516 00:30:08,750 --> 00:30:13,083 for one indication, say tocilizumab for rheumatoid arthritis 517 00:30:13,500 --> 00:30:16,041 and yeah have you you have the possibility of then 518 00:30:16,333 --> 00:30:20,500 using that in a different indication that that would be a repurposing opportunity. 519 00:30:20,500 --> 00:30:22,166 So for example in eczema. 520 00:30:23,125 --> 00:30:23,875 I guess it 521 00:30:23,875 --> 00:30:30,083 doesn't make sense to think about using an anti rheumatoid arthritis drug. 522 00:30:30,166 --> 00:30:30,458 Right. 523 00:30:30,458 --> 00:30:32,583 That's on market to treat eczema. 524 00:30:32,583 --> 00:30:35,541 That's just I mean. 525 00:30:35,541 --> 00:30:38,125 There's one in clinical trials. 526 00:30:38,125 --> 00:30:41,250 I mean, coming back to the cohorts, Cindy, I think also 527 00:30:41,750 --> 00:30:45,500 the fact that these cohorts there from different geographical places 528 00:30:46,333 --> 00:30:50,375 increase the possibility to illuminate, for example, biases on SNPs. 529 00:30:50,375 --> 00:30:51,083 Right. 530 00:30:51,375 --> 00:30:54,416 Did you have any discussion with the authors about that? 531 00:30:54,750 --> 00:30:57,416 Do they ever consider that the bias, geographical bias 532 00:30:57,583 --> 00:30:59,666 may influence their data? 533 00:30:59,666 --> 00:31:01,041 Can you comment on this? 534 00:31:01,041 --> 00:31:02,708 Yeah, it's a great question. 535 00:31:02,708 --> 00:31:05,791 So they primarily 536 00:31:05,791 --> 00:31:08,458 represent northern European populations. 537 00:31:08,916 --> 00:31:13,250 There were there was some representation of Asian populations in there, 538 00:31:14,458 --> 00:31:16,500 but not a not a lot. 539 00:31:16,500 --> 00:31:20,791 And I'm trying to remember, I don't think there was any African diaspora 540 00:31:21,208 --> 00:31:22,166 in this milestone 541 00:31:22,166 --> 00:31:25,416 paper in the subset of samples that they had in this milestone paper. 542 00:31:25,916 --> 00:31:29,416 So that's that's a you know, it's a blessing and a curse. 543 00:31:29,416 --> 00:31:30,333 Right. For them. 544 00:31:30,333 --> 00:31:35,708 It eases the analysis to your point for the opportunity to make discoveries 545 00:31:35,708 --> 00:31:40,041 because of diversity within the ancestry of our genomes. 546 00:31:40,041 --> 00:31:41,333 It’s a “miss”, right? 547 00:31:41,333 --> 00:31:45,291 And an enormous potential future opportunity, which I think 548 00:31:45,833 --> 00:31:50,125 is very exciting and very important for equity in health care. 549 00:31:50,333 --> 00:31:51,541 I mean, essential. 550 00:31:53,000 --> 00:31:54,958 So we have to start somewhere, though, right? 551 00:31:54,958 --> 00:31:59,250 So we start with the populations that we have. 552 00:31:59,250 --> 00:32:01,333 It's fascinating thinking about 553 00:32:02,291 --> 00:32:05,583 the 90 proteins, all the different things that discovered. 554 00:32:05,583 --> 00:32:08,958 Right, these 25 drug targets for. 555 00:32:08,958 --> 00:32:12,083 That explains why the pharma interests in the UK Biobank 556 00:32:12,583 --> 00:32:15,208 by doing the extrapolation. 557 00:32:15,208 --> 00:32:16,708 Have you done the extrapolation? 558 00:32:16,708 --> 00:32:19,791 How many drug targets they expect. Yeah. 559 00:32:19,791 --> 00:32:23,500 So with this, you know it’s around five and a half percent of the pQTLs 560 00:32:23,541 --> 00:32:26,666 discovered in Folkersen et al, converted 561 00:32:26,666 --> 00:32:29,750 to, you know, potentially causal. Interesting. 562 00:32:29,958 --> 00:32:34,250 So if we applied that same percentage which is lofty, right, that's 563 00:32:35,333 --> 00:32:37,333 is a lot of proteins and, 564 00:32:37,333 --> 00:32:42,291 and these 90 in Folkersen et al were well 565 00:32:43,291 --> 00:32:47,458 studied you know considering across 30,000 samples so you know I would 566 00:32:47,541 --> 00:32:51,500 I would expect maybe four, four and a half percent to maybe 567 00:32:51,500 --> 00:32:54,208 5% converting. 568 00:32:54,541 --> 00:32:58,166 In this initial set of proteins, I think to be a little conservative, 569 00:32:58,166 --> 00:33:01,458 you know, not not trying to be too bullish, but even with that, we're 570 00:33:01,458 --> 00:33:07,750 talking about potentially listing off causal markers to examine, to investigate 571 00:33:08,208 --> 00:33:11,500 potentially causal markers of, you know, around 500. 572 00:33:12,250 --> 00:33:13,666 So that. 573 00:33:13,666 --> 00:33:17,166 Five hundred potential drug targets.” 574 00:33:17,166 --> 00:33:19,291 Potential therapeutic targets. That's right. 575 00:33:19,333 --> 00:33:23,666 And to be fair, some of these might show up as potential therapeutic targets 576 00:33:24,041 --> 00:33:28,041 that would never be considered if they're in signaling pathways, for example. 577 00:33:28,250 --> 00:33:30,000 So so it's up to pharma. 578 00:33:30,000 --> 00:33:31,791 And certainly people that are 579 00:33:32,750 --> 00:33:33,833 are more versed in 580 00:33:33,833 --> 00:33:38,625 clinical trials and potential, you know, pathways 581 00:33:38,625 --> 00:33:43,041 for these and implications of side effects to then up score and down score these. 582 00:33:43,500 --> 00:33:46,041 But the exciting aspect of this 583 00:33:46,041 --> 00:33:49,125 is to have a systematic approach by which to do that, 584 00:33:49,250 --> 00:33:53,500 to actually make that list of 500 and then up score some and start programs. 585 00:33:53,500 --> 00:33:55,500 Because we like to say 586 00:33:55,500 --> 00:34:00,416 that clinical trials are twice as likely to be successful. 587 00:34:00,875 --> 00:34:03,791 If you go into that trial with genetic information 588 00:34:03,791 --> 00:34:06,958 that's certainly, you know, been published and we like to say that 589 00:34:08,083 --> 00:34:11,208 adding proteomic data, I'd really love to see 590 00:34:12,208 --> 00:34:15,458 what that means for our potential for 591 00:34:15,791 --> 00:34:18,958 for improving our ability to be successful in clinical trials. 592 00:34:19,250 --> 00:34:24,583 And these 13, I guess if you take those 500 targets (or potential drug targets) 593 00:34:24,583 --> 00:34:28,583 divided by 13 different pharma partners, that's like, what, 35 apiece. 594 00:34:28,791 --> 00:34:30,916 Yeah, that's right. That's a lot of programs. 595 00:34:30,958 --> 00:34:32,375 That's a lot of programs. 596 00:34:32,375 --> 00:34:35,458 I mean, that's going to be a wealth of data for them. 597 00:34:35,458 --> 00:34:39,958 Now, I understand why they would invest in such a project 598 00:34:41,041 --> 00:34:42,958 when what is the next step 599 00:34:42,958 --> 00:34:47,458 then in the UK Biobank project and how people find out more about it? 600 00:34:47,750 --> 00:34:48,708 Good question. Yeah. 601 00:34:48,708 --> 00:34:54,041 So the what I fully expect and I know of at least at least eight 602 00:34:54,041 --> 00:34:57,791 abstracts that have been submitted for ASHG this year. 603 00:34:57,791 --> 00:35:00,833 Now ASHG, American Society for Human Genetics, 604 00:35:00,833 --> 00:35:03,000 as I mentioned earlier, will be 605 00:35:04,041 --> 00:35:06,208 in Los Angeles in October. 606 00:35:06,208 --> 00:35:11,000 And so I know that those pharma partners and the researchers within those pharma 607 00:35:11,000 --> 00:35:14,958 partners are submitting abstracts to present there 608 00:35:15,291 --> 00:35:19,375 and I'm sure some of them will get oral or oral presentations. 609 00:35:19,375 --> 00:35:21,500 Many of them will get poster presentations. 610 00:35:21,500 --> 00:35:25,666 But I will be keeping a close eye on that and I will absolutely be there. 611 00:35:25,666 --> 00:35:28,041 And I think we should do a podcast episode. 612 00:35:28,208 --> 00:35:29,000 There you go. 613 00:35:29,000 --> 00:35:32,750 You will have a post ASHG. 614 00:35:32,750 --> 00:35:34,500 This is what I got out of it. 615 00:35:34,500 --> 00:35:35,750 That would be great. 616 00:35:35,750 --> 00:35:39,875 And maybe drag a few guests on if if we can. 617 00:35:39,875 --> 00:35:42,458 That's great. Yeah, that'd be great. Yeah. 618 00:35:42,750 --> 00:35:46,416 And as far as what I think is next, think they're going to be digging into these 619 00:35:46,916 --> 00:35:50,250 these correlations, 85% of them novel. 620 00:35:50,250 --> 00:35:53,916 So roughly 8000 novel relationships between genetic regions 621 00:35:53,916 --> 00:35:55,333 and protein levels. 622 00:35:55,333 --> 00:35:58,291 They're going to be looking into which of those are appear 623 00:35:58,291 --> 00:36:00,500 causal within certain diseases. 624 00:36:01,291 --> 00:36:03,125 Do you know when that will be available? 625 00:36:03,125 --> 00:36:05,166 Publicly-available data? 626 00:36:05,583 --> 00:36:08,541 How how scientists can have access to these? 627 00:36:08,541 --> 00:36:11,291 Is it a easy process or a difficult process to have access on that? 628 00:36:11,750 --> 00:36:12,041 Yeah. 629 00:36:12,041 --> 00:36:16,250 So as, as you probably know, Sarantis, but our listeners may not know the UK 630 00:36:16,250 --> 00:36:19,583 Biobank data through a data use agreement 631 00:36:20,541 --> 00:36:22,916 is, is broadly available. 632 00:36:22,916 --> 00:36:28,000 So this is one of the, the reasons there's so much use of those data as validation 633 00:36:28,000 --> 00:36:31,541 data and for discoveries with very clever 634 00:36:31,958 --> 00:36:34,875 informatics scientists and biologists to 635 00:36:35,833 --> 00:36:36,208 think of 636 00:36:36,208 --> 00:36:38,833 creative ways to use such a large dataset, 637 00:36:39,458 --> 00:36:43,041 the proteomics data, the first set of proteomics data. 638 00:36:43,041 --> 00:36:48,208 So the first 1500 proteins, the subject of the June bioRxiv paper 639 00:36:48,583 --> 00:36:50,958 Those data have been stated 640 00:36:51,208 --> 00:36:54,041 that they will be publicly available by the end of the year. 641 00:36:54,958 --> 00:36:59,166 So I expect, you know, by October at ASHG we'll know better. 642 00:36:59,166 --> 00:37:01,416 The timing for that, 643 00:37:01,416 --> 00:37:05,041 yeah, those pharma partners, of course have had access to those data 644 00:37:05,041 --> 00:37:08,375 as they should, which is why they were able to publish that, 645 00:37:09,125 --> 00:37:11,416 that paper so quickly. 646 00:37:11,416 --> 00:37:16,375 And so the next tranche of data for the full 3000 proteins. 647 00:37:16,375 --> 00:37:18,083 And can I just say, you know, 648 00:37:18,083 --> 00:37:22,041 you see what's possible with 90 proteins and Folkersen et al. 649 00:37:22,041 --> 00:37:26,125 Imagine what's possible, you know, with 30,000 650 00:37:26,125 --> 00:37:30,125 proteins, 3000 proteins and 54,000 individuals. 651 00:37:30,125 --> 00:37:33,833 That's a lot of power to deduct relationships between proteins and 652 00:37:34,166 --> 00:37:37,000 and many proteins that really just haven't had assays 653 00:37:38,333 --> 00:37:39,875 for, for examining them. 654 00:37:39,875 --> 00:37:42,125 So just such a such an opportunity for discovery. 655 00:37:42,958 --> 00:37:47,000 We touched upon yeah, we touched upon the enormous investment 656 00:37:47,000 --> 00:37:51,000 made to-date to collect these 500,000 samples 657 00:37:51,000 --> 00:37:51,458 That's right. 658 00:37:51,458 --> 00:37:54,125 And to follow up and all those like genetics. 659 00:37:54,125 --> 00:37:54,666 Yeah. 660 00:37:54,666 --> 00:37:57,791 The whole genome, whole exome data on all these individuals 661 00:37:58,208 --> 00:38:03,208 and then now overlaying empowering the genomics with the proteomics. 662 00:38:03,208 --> 00:38:06,875 It's as if we're a part of something that is the next 663 00:38:06,875 --> 00:38:09,916 big thing in genetics is proteomics. 664 00:38:09,916 --> 00:38:12,416 I think it's you know, and when you think about the 665 00:38:13,458 --> 00:38:15,500 the central dogma of biology, right? 666 00:38:15,500 --> 00:38:16,583 You've got DNA. 667 00:38:16,583 --> 00:38:19,708 RNA, we've done a great job of looking at DNA. 668 00:38:19,708 --> 00:38:23,458 RNA has been our proxy for time biology for a long time because it was 669 00:38:23,458 --> 00:38:28,583 it was available to to look at with sequencing technologies. 670 00:38:28,583 --> 00:38:33,041 In fact you and I Dale, I think have talked about how the RNA-Seq 671 00:38:33,083 --> 00:38:36,250 and the ability to do what we call “digital gene expression” sold 672 00:38:36,250 --> 00:38:38,375 many of those initial instruments 673 00:38:39,333 --> 00:38:42,291 that were, you know, next generation sequencing instruments. 674 00:38:42,958 --> 00:38:45,875 But now we have this this ability to measure 675 00:38:46,250 --> 00:38:49,875 proteins directly in a in a very scalable way. 676 00:38:50,250 --> 00:38:53,125 And I am excited, as you 677 00:38:53,125 --> 00:38:55,208 know, about this capability, 678 00:38:56,250 --> 00:39:00,333 but it's really the researchers and what they can do with it 679 00:39:00,333 --> 00:39:05,083 that will tell us the true potential of this. Super. 680 00:39:05,333 --> 00:39:06,666 Well. Thank you, Cindy, 681 00:39:06,666 --> 00:39:10,041 for sharing your thoughts on empowering genomics with proteomics. 682 00:39:10,458 --> 00:39:12,208 And we'll see you soon. 683 00:39:12,208 --> 00:39:12,958 That was great. 684 00:39:12,958 --> 00:39:15,083 Thank you very much. 685 00:39:19,041 --> 00:39:22,833 Thank you for listening to the Proteomics in Proximity podcast brought to you 686 00:39:22,833 --> 00:39:26,833 by Olink Proteomics. To contact the hosts or for further information 687 00:39:27,083 --> 00:39:31,750 simply email: info@olink.com.