Dr Sarah Beecroft from Pawsey Supercomputing Research Centre
“it’s fun. Like you, you know, uncovering kind of the deeper insights, I think, which is quite interesting as well. You’re getting closer to the bigger story or the deeper story if you want to look at it that way. ”
“you kind of need to like get to know your data set. So you’ve got to, you know, get in there and tool around with it. Get your hands dirty, so to speak and explore it, which is, it’s like a, it’s kind of like an adventure”
“But being able to confidently say like actually the evidence shows this and we found this unexpected thing. It’s just like so exciting. Yeah, really, I just love it.”
“I like puzzles. I like figuring things out. I really always want to understand the why of these types of things. And I love biology. So it’s like, puzzle biology. Well, what more could you want?”
TRANSCRIPT
Linda: Welcome back to another episode of Make Me Data Literate. I’ve been wanting to interview this guest for a while, so I’m very excited to bring you Dr. Sarah Beecroft. Welcome, Sarah.
Sarah: Yeah, thanks. Thanks for having me. So, let’s start off with, who are you and what do you do? Well, as you said, I’m Dr. Sarah Beecroft. My job now is working at the Pawsey Supercomputing Research Centre as a super computing specialist. So, this means that I help researchers and scientists from all around Australia use our supercomputer. So, this might mean that I give them training. It might mean I’m troubleshooting their code. It might mean that I’m working in the background to implement new features that researchers want. But before doing this, I was actually a biologist. So, I did a human biology degree, which feels very disconnected from where I am now, working in computing. So, yeah, I did the bio degree. I didn’t do any coding there. I did an honours at the University of Western Australia and picked up a little bit of coding. And I realised that I actually really liked data analysis and statistics, which I wasn’t expecting. And then did a PhD in rare disease, or like rare neurogenetic disease, gene discovery. So, finding out what genes are making people have these really terrible diseases. And that is actually pretty computationally heavy because to analyse all the genetic data, you need to use computers. So, I was what I would jokingly call a soggy scientist because I could do both wet and dry lab.
Yeah, I could do the data analysis side and I did that and I liked that part the most. But I could also go into the lab and chop up a piece of someone’s muscle from a biopsy, extract the protein out of it and do a western blot to analyse the abundance and the size of the protein. I could grow cells in a dish from people that have passed away. You take fibroblasts while they’re alive and often they don’t survive. And you end up growing these cells trying to figure out what happened and making, transfecting other different types of cells to grow glow green and doing confocal microscopy.
But I left all that behind, all the wet lab stuff to focus on my one true love, which is data analysis and getting into the pretty computational aspect of things, which I realised actually I really, really like. So, that’s something I really enjoy doing at Pawsey.
Linda: That is so cool. You can do all of the things.
Sarah: Yeah, I guess so. Arguably, I wasn’t amazing, I have to say it, but well, stuff. Cell culture was always a struggle for me because you have to have incredible focus for a long time so you don’t contaminate your cells. You can’t put your hand over the open dish or something because then, like, that’s bad practice and you might get contamination. So, that wasn’t so good.
But the other parts, like the Western blots I got pretty good at and growing, doing like really finicky, delicate things. We were growing cells on these really thin glass cover slips and then you had to take out, do all these things to these glass cover slips that were in the bottom of these wells, take them out with a pair of tweezers and glue them basically onto the glass slide so you could image them under the microscope. And so doing that with your pair of tweezers, I’m proud to say I don’t think I broke any cover slips, which, yeah, my supervisor couldn’t manage it, she broke loads, so I was so proud of myself.
Linda: That’s incredible. I can confidently say I do not have the fine motor skills for that kind of activity.
Sarah: Yeah, it’s not something I realised I could do. You find out all these funny things about yourself when you’re put in these weird situations, and you’re like pretty mediocre, like gross motor skills, like I’m quite clumsy sometimes, but fine motor skills, top notch apparently. So, hooray for that, because that’s something that you use all the time in your daily life, for sure.
Linda: [laughter]
So, what did you have to learn to do what you’re doing now? You know, we’ve already sort of got a hint of that, but I’m fascinated by the way so many people get into computing, and especially high performance computing, HPC, via something else entirely, and then they teach themselves some, you know, tech skills, and suddenly they’re in a whole new industry.
Sarah: Yeah, I mean, it’s pretty wild, and I totally agree with you. Most people don’t follow a linear career trajectory in these types of fields. So, I haven’t met pretty much anyone in a STEM field that has, well actually no, that’s not true, I went to school with one person, and she went to what she called ‘Nerd Camp’ in year 10, and she decided at that point that she was going to become a chemical engineer, and I’ve lost touch with her, but like I’m pretty sure she’s still a chemical engineer now, so, you know, she had a linear career path, and that’s actually great, but pretty much everyone else, I’ve met has had this kind of popcorn one thing going to the next.
So for me, yeah, the human biology, I had no idea about coding, I didn’t realise that would actually be something I could do or be interested in. I fell into it by accident, as a lot of people do, trying to solve an unmet need within our lab. So, with all of this genetic data, we had to do all this like post-processing. So you get the sequencing data back, and that’s cool, but it’s not like a biological insight yet, it’s just something, you know, you have to do a lot of processing and filtration to get something useful out of it that you can then like scientifically analyse.
And so, I picked up a bit of bash coding there to improve our data processes, and so actually implementing a bash for loop was like a big innovation in the lab at the time. Wow. Yeah, because we just didn’t have the skill set, like it was a wet lab originally, and then it’s kind of slowly transitioning to be wet and dry. So, yeah, I kind of taught myself that, and it was a pretty slow process, I have to say, it took a while to get going, and then I kind of kept building on that throughout my PhD.
But the main thing at Pawsey is they were looking for someone that can do this translational piece, I suppose. So, something I had been doing a lot was doing like training and like networking events, so like, you know, I saw an unmet need in the community, so I started doing those events, and Pawsey was like, oh, this is the type of person that we need, because you can talk to people, but you also get biologists and you can do a bit of the computing stuff. And so, actually, when I started at Pawsey, I had not really run a bunch of jobs on the supercomputer, so I had to learn, you know, all the scheduling stuff, and basically all the inner workings of HPC I learned on the job, which has been really cool and really interesting, but also bad for my imposter syndrome, because it’s like, oh, these other people, they know how to do things in FORTRAN, and like, I don’t know that, and oh, but I think what I bring is, you know, it’s like the bigger picture thinking, which is what we’ve needed to support life sciences in HPC, where there has traditionally been a disconnect. But now I can do a lot of the technical stuff quite well, so I’m happy about that.
Linda: I think that confirms what I said on Einstein and GoGo a couple of weeks ago, which is that the tech is easy, it’s the people that are hard. And, you know, you hit the wall with tech and get frustrated, and it can feel insurmountable. But once you’ve got it, you’ve got it, whereas you can understand a person one day and be completely flummoxed by them the next. So, you know, it’s much more complicated, really, the people side.
Sarah: Yeah, I agree with that. Not to downplay the technical skills of the people that I work with, but also once you get up to the higher decision-making things, like, what should we do? Where should we go? How are we going to get the money from the government to do this? That’s all people. And it’s less of a science and more of an art, which is really hard. Like, some people, they make it look easy, but if you’re not one of them, you just… Yeah, it can be very tricky. So, I think it’s important, like you say, to have this recognition that those skills are important, for sure.
Linda: The CEO of Pawsey is going to love this conversation, seeing as he’s a literature major.
Sarah: Yeah, well, we’ve chatted before about that, like with him, or, like, you know, corridor conversation type of thing, and it is really important. And, you know, for someone like him, like, he needs to be really good at this people part, and he is. It doesn’t matter if he can code in bash or Fortran or whatever, like, it doesn’t… It’s irrelevant to his skillset. What we need is someone like him to go out and do the people and the strategy.
Linda: Yes, absolutely. But this man we’re talking about, Mark Stickells, is also very open about the fact that he also has imposter syndrome, and he feels, you know, he’s very conscious of his lack of technical skills, whereas, you know, he could learn them if he wanted to, but he doesn’t have the time or the need. And that comes back to this idea that is, you know, again, confirmed by your story, which is, it’s much easier to learn technical skills when you have a problem to solve, when you have a reason to learn this particular skill, and, you know, you’ve got this unmet need, and you’re like, well, hey, if I could do this with tech, I could do it faster or more efficiently or more accurately. Suddenly, you’ve got a reason to learn it, and I think that’s been missing from our coding education and also our maths and statistics education. Did you learn any code at school?
Sarah: No. So I’m a millennial, I’m 35 now. So no, it wasn’t a thing. Like, we did have computing as a class, but I was like, I actually don’t know what they learned in there, but I was also like, that’s not relevant to me, like, ew.
Linda: Right?? Yes. Yes, people still think that. It’s not relevant. It’s not interesting. You’ve got to make it relevant. You’re just validating my work here. So smug mode. Is there one thing that you wish everybody knew about data, something that you just go, oh, it would change everything if people just got this thing?
Sarah: So yeah, that is a big question. I think something I would like to see people have is better foundational understanding of kind of how to understand numbers. And figures that are presented to you, you know, by your bank or in the media or these types of things. Because it’s so easy to misrepresent data and facts. And so it looks legit, but it’s really not. And so I see a lot of that around. I mean, it’s so common. I would hope people could be empowered to see through these dodgy kind of metrics. I think that would be great.
And also financially as well, because I mean, that’s the type of data. I think it’s empowering if you can be financially literate. And it took me a while to learn that myself. Like when I came out of school, I was terrible with money and now I’m a lot better, thank God. But yeah, those are two things I think would really help the everyday person.
Linda: Yeah, I completely agree. And, you know, I think I’m just going to cut and paste this entire podcast as a validation for the existence of the Australian Data Science Education Institute. Like, this is, you know, this is the thing, putting these skills into the context of stuff that actually matters that will impact your life. And, you know, when somebody shows you a graph and there’s no zero on the scale, what are they trying to do? You know, what are they trying to make you think? What’s the story here? What are the worst data mistakes that you’ve seen?
Sarah: So I guess this is coming more from a professional, science-y type of bent, but you definitely see some people that maybe don’t have a great understanding of basic statistics. And when you’re designing a scientific experiment, you want to make sure that you have thought about this type of thing. And it’s a really, really, really common complaint from, you know, bioinformaticians, people that are doing this data analysis, and they’re meant to be the ones with a strong statistical understanding that they will have someone come in and be like, hey, I’ve got like two replicates. Can you like do statistics on this? It’s like, dude, I can’t do anything with two of anything. Like, that is not enough for science. Do you think you could have like spoken to me about this like six months ago before you spent all this money on a stupid thing? And yeah, it’s just like, what are you doing? Please, like, either learn the statistics or come talk to me first. So I think the basic lack of statistical analysis is really concerning. And that’s quite a widespread problem.
You know, like there was, I did a little bit of statistics in my undergrad because I did an epidemiology unit. And so that was actually really good. But it wasn’t like a core requirement or anything. So you could have easily gone through and not done that. And then I started a masters, which I didn’t finish. And part of that included some like a unit of statistics as coursework. And that was actually so, so useful. And I’m really glad I did that. But most scientists don’t actually get the hardcore statistical training. And it’s a huge gap. It’s a huge gap. And yeah, I just think it’s a big problem. Like, you shouldn’t have only one person in the team who doesn’t get listened to that’s doing your statistics.
Linda: So many statisticians who are listening to this are going, oh, God, yes. Hitting their heads on the desk. I’m astounded by the fact. And I noticed this when I started ADSEI that it’s still possible to do a science degree and not cover data at all. And the data management and analysis and stuff that I see scientists doing really is like horrifically,
Sarah: bad
Linda: yeah, shocking. It’s just shocking. And it’s not deliberate. It’s just they don’t know any better because it’s not part of the training. And I don’t think we can keep doing that.
Sarah: It’s not really enough. And the other thing is people finding it really scary. Like, I like statistics. And I think my dream job would actually be to be like a data scientist. If I could do it all again, maybe I would do that. I probably wouldn’t. I like where I am. But you know, like, that sounds fun to me. So I’m weird in that way. That’s, you know, this idea, you know, I see from other people in the lab who are really good at, you know, what they were good at and being like, okay, but you know, we need to think about the statistics and like, oh, no, oh, that’s, I don’t want to do that. It’s not, it is a little bit scary to start with, but it’s, it’s okay. Like you managed to learn all this other really difficult stuff. You can learn this. So maybe it needs to be taught differently. I don’t know.
Linda: Yep. Yeah, absolutely.
Sarah: And what’s the difference between a p-value and a confidence interval? Like the p-value is… we’re always chasing the small p-value to show that it’s true. Like, well, no, that’s actually a lot more complicated than that. What is your confidence interval? Like, what’s the effect size, blah, blah, blah, blah, blah, all these other things. So yeah, I guess, you know, you have to work with the data that you have and help it become, it’s almost like, you know, extracting like a sculpture out of a piece of stone or something. Like there’s a story in there is trying to tell you, you just have to work with it to get it out. And once you find it, it can be so interesting and kind of beautiful in a way. I don’t know if that’s a strange perspective to have, but whatever.
Linda: No, no, it absolutely speaks to me. And it’s one of the things, you know, I work with data a lot more than I ever have before in this job. And I’m just like, I get so excited when I see a data set and I think, oh, I could figure out this thing, you know, or, oh, what can it tell me? And what can I find out? And it’s just, it’s such a thrill. And sometimes you get, you know, such insight that you didn’t have before. And, you know, it’s a really powerful feeling. And I’m sad that we don’t offer that to kids the way we currently teach it. You know, I looked, I was looking, I’m giving some feedback to ACARA on the senior secondary math curriculum and I was looking at some of the resources around for the stats section of math methods, in particular the probability section. And all of the resources are coin tosses and, you know, taking colored urns out of balls, sorry, colored balls out of urns. And I’m just like, why, why would they care? Like students, like, agghh, … we could be doing so much better than this. You know, what about the odds of getting through to the switchboard when you’re trying to buy Taylor Swift tickets?
Sarah: That’s a hot topic.
Linda: Right? Like, you could make it so much more interesting while teaching the same content and so much more relevant and give them actually a reason to care about it. Sorry, You may have just pushed a button there.
Sarah: No, I mean, that’s great. And it’s interesting to me.
Linda: It’s beautiful if you know what it can do.
Sarah: Yeah. And I don’t, you don’t get to find that out until you’re like really, really, really down the rabbit hole. And I never would have like, I actually wasn’t particularly great at math at school because a lot of it is a arithmetic, right? Oh, as in like, you know, even in advanced math, you’re still like, they want you to mentally work out X things. And, you know, if you put the number in the wrong place and the whole thing’s broken, which I was always really bad at, making, making a lot of these quote unquote, “Careless mistakes.” That’s so frustrating. And, but once you’re actually doing like legit statistics, you don’t do any of the maths yourself. You understand the concepts and you’re thinking in a different way. And but you don’t get to do the fun stuff until you’ve suffered all these other things first.
Linda: Yes.
Sarah: And I always thought like, I’d never like statistics because I didn’t really love maths, but it’s not true at all. Which I was really surprised about, like, I fell into it by accident because it was applied, basically.
Linda: I literally have a friend who is a biostatistician who says the same thing. We’ll get her on the podcast eventually. But, but yeah, like, she, she hated maths at school. She now has a PhD in stats and she’s an associate professor. And she’s like, yeah,
Sarah: Wow. That’s amazing.
Linda: It really says something about the way we’re teaching it and the experience of it in schools that we’re really missing the mark quite severely. Wow. Yeah. All right, got distracted from the questions there, as I do. Have you ever seen data deliberately misused and how do you spot it?
Sarah: Oh, man, I mean, it’s kind of all the time, I think, even with these like, by now, pay later type things. You know, they’re not upfront with the what is it really costing you or like, sure, you can have this mortgage, but you actually work out, you know, how much interest you’re paying over the decades and it’s more than the house is worth.
So, I’m not saying that, you know, banks are misrepresenting … like the information is is there, but it’s, yeah, it’s hard to, I think, get the full picture of what does it mean for you as an individual, but also things like, you know, we had an election relatively recently in Australia. And I’m sure they would be, if you wanted to pull out some examples, there was lots of things about, can we afford to implement this nuclear power plan? And, you know, what is the real cost of that and it’s just so hard to truly know to cut through all the spin and say like, well, what is the most economically sensible solution, if not environmentally like that’s a different question, I suppose. How can you spot it? I mean, it’s so difficult.
I guess you need to be asking questions of like, you know, is there more information here than what the person is telling me have I read all the terms and conditions. Yeah, I don’t think I have like a single tip for like, oh, you need to watch out for the error bars type of thing. Unfortunately, I’m sure other people or yourself have talked about it in other episodes, but I guess be suspicious. Trust no one.
Linda: It’s that the challenge then is, and I like to call it like rational skepticism, you need to be skeptical, but you need to be rationally skeptical and you know, not just like wholesale “science is all lies”.
Sarah: Oh yeah,
Linda: there’s a fine line between being skeptical to the point of conspiracy theories and being rationally skeptical and actually being able to evaluate the evidence, which is a big part of like a big part of what I do. Yeah, try to train kids and teachers to, you know, be on the correct side of that line, I guess.
Sarah: Yeah, absolutely. Like, you know, I guess it’s more seeking to understand rather than being overly skeptical. Like if someone tells you, like, oh, I had this really great holiday in Bali and I went to this place and I did this thing and you’re like, okay, well, that sounds nice, but I know that you’re a party animal and I like to be in bed by nine. So you would kind of like, you listen to what they’re saying, but you’re like, okay, but like, you know, what’s the bigger picture? What were you doing there? Would I like that? Is this, you know, you take it on board and you listen, but you’re also like, well, is there more to the story? Not like everybody is out to get me with data.
Linda: It’s like my grandfather lived 100 years and smoked, you know, two packs of cigarettes a day. Well, okay.
Sarah: I’m sure he did, but that is not how it is for most people. And it would be really interesting to study him and figure out how he managed to tamp down all the mutations that should have caused him cancer and he didn’t. Like, he’s an outlier and that’s biologically really interesting, but it doesn’t mean that you should smoke or vape.
Sarah: Yes, there was a whole interesting thing that came up recently about the longevity studies, you know, these kind of these little communities that have people who’ve lived unexpectedly long and a lot of them turn out to be anomalies of data collection.
Linda: Oh, really? Yeah, so one of them, I can’t remember the area, but one of these little communities that turned out that they just weren’t reporting the deaths because you’d still collect like, I don’t know, some form of benefit or food or something like, you know, so there were things like that where, or their birth dates were a little fudged or, you know, things where you actually look into. When you actually look into it, it turns out it’s not real.
Sarah: Wow, that’s amazing and kind of sad. But that’s a very true concept as well, I would say for bioinformatics and science in general is like garbage in garbage out . If your input data is garbage, don’t be expecting to get something really good at the other end. So that has to be pretty stringent.
Linda: It’s always that question then of how good is your data and also the question that I always tell people to start a project with which is what’s wrong with my data. Because there’s no such thing as a perfect data set. So the question is what’s wrong with it, how wrong is it and is it good enough for my purposes. And you know what’s good enough for my purposes might not be good enough for Sarah’s purposes, depending on what it is that you’re trying to do with that data. And I think a lot of a lot of issues that we see coming up as a result of data science are people trying to do things with data that it’s not up to. You know, it’s not … that data is not good enough for that purpose. So it’s not … doesn’t answer that question or it doesn’t have the level of accuracy required. It’s particularly problem when you get, you know, instruments and things that report to four decimal places, but actually don’t have the level of precision to get to four decimal places so that, you know, the last three of those decimal places are completely fabricated.
Sarah: Wow, like rounding error or something.
Linda: Yeah,
Sarah: that’s amazing. I really like that as an approach. I have to say like what’s wrong with my data. Because it forces you to think about the limitations which there always are. And if you know what they are, well, that’s that might be fine because you can adjust for it statistically or you can get around it or you can just say, these are the conclusions I can confidently make. These are the conclusions that I could tentatively make but I’m not really sure about and these are the things I cannot say. And like, yeah, that’s really strong. That’s a really strong approach. That’s really awesome. I promise Linda didn’t pay me to say that.
Linda: This whole episode is just such good advertising. We did not set this up at all. Yeah, it’s, I think it’s incredibly important and being able to say what you can’t do with the data is a skill that we don’t teach at all. And I’m trying to build that into my projects always to go, you know, where is the line here? I mean, I’m fascinated. I’m working on losing weight at the moment and I can step on the scales, off again, and then on and get up to three kilos difference.
Sarah: Wow.
Linda: These are nice new scales and they have a digital readout and we tend to kind of assume or just kind of believe that digital scales are more accurate than analog scales with the, you know, the wobbly needle, but they’re not. And the variance between ratings can be, so I’ve got a spreadsheet now and I measure it three times every time.
Sarah: Wow.
Linda: You know, the variance can be zero or the variance can be four kilos. I have not figured out why, like, and there may not be a why, like it’s just that this is, they’re not accurate, you know, they’re not, they’re not that accurate. But we assume that they are, you know, and sometimes we would get on the scales one day and go, oh, you know, 300 grams more than it was yesterday. That’s noise. That’s not a real result. 300 grams is noise in the face of that level of inaccuracy. So I think understanding that is a really important skill.
Sarah: Yeah, wow, I hadn’t noticed that, which is really interesting. And you might have people thinking like, oh, it’s gone up a bit today. Oh, no, what did I do wrong type of thing? I have a friend that’s a dietician now. And one of the things she does say is like, you know, if you are measuring yourself every day, like that’s fine, but don’t take it too much to heart because like maybe you just need, you know, you drank a bit more water yesterday or something. So, you know, don’t worry about it. But that’s actually missing this potential instrument inaccuracy, which, yeah, I don’t know that her or I have thought about.
Linda: Yeah. Well, if you have scales available to you, I invite you to try to step on it off a few times and do it a few different days and say what kind of results you get. It’s really interesting.
Sarah: I think I will. I do have scales. I’m going to try it. I have to now. I mean, I just, you know, love that type of thing.
Linda: You can tell me whether it was whether you were getting big levels of variance or not. I’ll be interested. What’s the first question you ask when you look at graphs in the media?
Sarah: Oh, yeah. <sighs> Good question. I mean, I think the biggest thing is who is making this graph and do they tend to be reliable in general? So, for example, if I see a graph from the Guardian or ABC, I’m a lot more likely to be like, yeah, that’s probably legit. As opposed to, I don’t know, a graph from someone trying to sell me something or from a political party. I never really trust those. I have like – no offense to any political party. Just have to, they are trying to make a point, right? Just about their party.
So, yeah, like, what is the story it’s trying to tell? And what information have they excluded from this graph as well? And, you know, do, have they chosen a format that makes sense? So, sometimes, you know, you see something like, well, there shouldn’t really be a bar graph that should be represented in this other way to give you a bit more information.
If I’m looking at something a bit more sciencey, you know, I want to see. You know, is there a line of best fit or the error bars or like how well is this explained and how well does, how strongly supported are their conclusions from this graph and from the data set, which I guess isn’t usually in the media. But, yeah, what are they not showing? I think it’s the biggest question.
Linda: Yeah, that’s a really important one. And there’s been some interesting studies that show you that people believe science reporting more if there are graphs in the article.
Sarah: Oh, wow.
Linda: Even if the graphs are completely irrelevant to the story, they did some great studies, you know, putting in random, you know, like, not just graphs but pictures, you know, they’ll put in pictures of the brain in a story and, you know, the pictures don’t give any information that’s relevant to the story, but people believe it more. And I think it’s difficult to train yourself to ask those skeptical questions. But once you start doing it, you can’t stop. It’s a little bit like when I was teaching my year 11s usability, they come to me and go, you have ruined me. I can’t even use a door now without critiquing the usability of it. Websites are a nightmare. I love that because they’re thinking about it and it’s the same with data. You know, you once you start thinking about it and you start asking those questions, you can’t stop. And that’s, that’s when you become a really potent force, hopefully for good.
Sarah: Yeah, and it’s fun. Like you, you know, uncovering kind of the deeper insights, I think, which is quite interesting as well. You’re getting closer to the bigger story or the deeper story if you want to look at it that way.
Linda: Yes, absolutely. It’s an opportunity.
Sarah: Yeah, exactly.
Linda: What excites you about data?
Sarah: Oh, yeah. I mean, I think what we just talked about earlier, which is finding the story that it’s trying to tell you. I really, yeah, I really like that. Like, you know, I really actually used to enjoy sitting down with a data set and be like, what have you got for me today type of thing? I remember I also used to say like, oh, you kind of need to like get to know your data set. So you’ve got to, you know, get in there and tool around with it. Get your hands dirty, so to speak and explore it, which is, it’s like a, it’s kind of like an adventure. So I used to find that really fun. And yeah, you, when you do get like some kind of insight out of it, it’s just so exciting. And like I did this study, which was 2,249. I wish there was just one more person. I don’t know why that bothers me so much. 2,249 patients who we did put through this gene testing panel in like a diagnostic sense. And so I was analyzing all of the patients. What disease did they have? What mutation in what gene? What was the diagnostic success rate from different age groups, from different disease groups? You know, who’s most likely to get diagnosed? Who’s least likely to get diagnosed? Where are the gaps? It was so fun. The data cleaning was horrendous, but once I got past that, it was really great. And yeah, I was just like, this is so cool because I can confidently say now, like we had kind of, you know, had the vibe in the lab of like this, you know, kids are more likely to get diagnosed. We know that. And, you know, these disease groups have a better diagnostic success rate. These other ones, it’s a bit more garbage. But being able to confidently say like actually the evidence shows this and we found this unexpected thing. It’s just like so exciting. Yeah, really, I just love it. I don’t know. Why? Why? I don’t, I guess it’s like a puzzle. Like I like puzzles. I like figuring things out. I really always want to understand the why of these types of things. And I love biology. So it’s like, puzzle biology. Well, what more could you want?
Linda: That’s awesome. I love that. That’s a beautiful note to end on. I think that whole puzzle thing, the whole, you know, the potential to figure stuff out as it’s a gift. s
Sarah: Yeah, I mean, it’s kind of a privilege to get to do that type of work, which I think is why a lot of academics and researchers put up with job insecurity and stress and short term contracts and an uncertain funding environment. They put up with it because they just love the work so much. It’s like addictive in a way. Yes. So, and apparently with mice looking at addiction studies, if you give the mice, so they like push on a lever and then they get some kind of reward, sugar water or something. I can’t remember. If you give them the reward intermittently when they do the thing, they’re more likely to keep trying. And I feel like that’s how it is with research. Like, oh, this one paid off.
Linda: Oh, yeah. Simple creatures, researchers.
Sarah: Yeah, they’re after the high. And the intellectual fascination. So, you know, and I did really love it, but I just couldn’t handle the insecurity and I was a bit burnt out. So working at Pawsey is much more stable and I still get to do intellectually interesting things, but with a lot less of the stress. But I look back very fondly on my time in research and, you know, the highs are so high. The excitement is really hard to beat, I would say.
Linda: Yeah. Yeah. That’s really cool. Thank you so much. This has been the best conversation. Yeah, I’m going to have to use it as advertising.
Sarah: How funny is that? I really had no idea that it would turn out that way. Yeah, it’s been really good. And it’s been really interesting to hear more about what you’re doing and just to nerd out. So, yeah, thanks. Yeah, gotta love a good nerd session. Thank you so much. [BLANK_AUDIO]
