Transcript#

This transcript was generated automatically and may contain errors.

Hey there, welcome to the Paws at Data Science Hangout. I'm Libby Herron, and this is a recording of our weekly community call that happens every Thursday at 12pm US Eastern Time. If you are not joining us live, you miss out on the amazing chat that's going on. So find the link in the description where you can add our call to your calendar and come hang out with the most supportive, friendly, and funny data community you'll ever experience.

Can't wait to see you there. I'm really excited to introduce our featured leader today. It's Chris Gates, Director of the Bioinformatics Core at University of Michigan. Chris, it's so nice to have you here today. I would love it if you could introduce yourself, tell us a little bit about what you do, and I will also ask you what you like to do for fun.

That's great. Thank you, Libby. It's great to be here. So yeah, I work at the Bioinformatics Core at the University of Michigan. Let me just explain that. Bioinformatics means different things to different people, so it's useful to unpack that a little bit. For us, bioinformatics lives at the intersection of three different domains. One is molecular biology, which you can think of for our purposes as DNA, genes, proteins, but also statistics, and finally computer science and software engineering, where those two overlap, AI now.

Really, what we do at the Bioinformatics Core is that we work with biological researchers, and they come to us with sequence data or proteomics data, and we help them identify patterns in DNA, then try to help them interpret what those patterns mean. So we do that a couple different ways. One way is that the algorithms that people use to analyze DNA are constantly changing, so we're always hip-deep in literature review to try to figure out what the best practices are and apply those and also share those out to the research community. Another way that we help people is that we have a series of workshops that are focused on bioinformatic analysis, and the deal there is that a lot of bioinformaticians, they might have some computational chops, but it's a steep learning curve, so we try to give them an intro to some of the basics there so they have more independence.

Then the final way we help them out is that we will do analyses on their behalf, and that happens a lot because biologists come to us and they have a deep expertise in their domain, so they really know a lot about diabetes or the human retina or adenocarcinoma, but they don't necessarily know all of the idiosyncrasies behind the informatics and all of, frankly, kind of strange data types and algorithms that can be a little bit wonky, and even if all those things work, they typically don't have the deeper training in how do you wire this up and execute it at scale in a computational reproducible way, so we help them with that.

I think that's probably raised a few questions just by itself, but that was a two-parter, absolutely.

Yeah, well, the other question that I was going to ask is what you like to do for fun, but now I'm dying because people in the chat are asking why there's an NBC logo in the bottom of your screen.

Oh, sorry, this is homage to Ron Swanson's office. This is not actually my office. Okay, but, okay, sorry. I should take that out, but yeah. Here, you can get the whole shot. You can take a snapshot and then use it for your own. There you go.

We figured it out. Somebody was like, oh, my God, that's Ron Swanson's office, and as soon as someone said that, I saw the bacon and egg, like the breakfast picture, and I was like, oh, my gosh, yes, but we will get the crazy train back on track for one second here.

Sorry, no, no, sorry. I didn't mean to distract everybody.

That's so good.

I should have a boring science lab in back of me, but no, sorry.

Absolutely not. Never change, but that brings us to what does Chris Gates like to do for fun?

This is such a weird thing. I have off and on for many years played drums in various groups and venues, but around last Thanksgiving, it's a long story, but the short version is somebody turned me on to steel pans, steel drums, sometimes people call them, and my wife and I just kind of went all in, and so now we have these three steel drums in our house, and whenever I can get any time at all, I'm on that steel drum and practicing, and now we're actually going to perform in an open mic coming up in a couple of weeks, so it really got away from us, but it's super fun. Yeah, so that's what I do.

Music gets away from you. It's a very good thing. Your house is probably very, very fun all the time. I am married to a drummer, and so I know the noise and chaos and fun that comes with that. Okay, well, we have a couple of questions in the Slido, and hilariously, they're both from Noor. Noor always gets us started with the best questions, so I will highlight one of Noor's questions and stick it in the chat here, and if we could have Noor unmute and ask this question, that'd be amazing. Hey, Noor.

Career path into bioinformatics cores

Hi, my name is Noor. I'm also a bioinformatician. You've got to keep me honest. I also understand the dilemma we often have of explaining what we do. I often say it's like, I'm like a biological data scientist, until I talk to other bioinformaticians. They're like, hi, I don't do protein. I thought it was kind of boring, so yeah, that's on me. I'm a microbial genomic, like microbiome and stuff, so that's my gen, but everyone's like humans, and I'm like, I guess I'll do them, but anyway, the question that I have or I'm allowed to ask, at least, is that my dream job is actually your job. How did you sort of get into that avenue, or how do you basically go to the university and be like, hi, I would like to just touch all your data and help you, and I'll do it for a reasonable, cheap fee, essentially, or hire me, but yeah, how did you get along that avenue, per se?

Yeah, that's a great question. I said at the outset, there's the notion that bioinformatics sits at the intersection of molecular biology and statistics and computer science and software engineering, and I feel like that's a useful kind of paradigm, if you imagine that three-circle Venn diagram, that in general, not a lot of bioinformaticians just kind of emerge from the center of that Venn diagram. What they do is they come in, and a lot of times, bioinformaticians will come in from the biology side, so that they're on the bench, they're working, they're running experiments, and then at a certain point, they start getting involved in some of the analysis, and that just kind of continues to build, to a point where they're like, I would rather be working on the computer than working on the bench, but that's not the only way to get there. There are people that are coming in from the stats side. I personally actually came in from the software engineering side, which is, it is not a well-traveled trailhead, but yeah, the short version of this is that in 2007, I hooked up with a biotech. It was building a giant data mining platform based on microarray data, and I just kind of got the bug for the molecular biology. I couldn't get enough molecular biology, and I was a trade software engineer at the time, and they were like, why do you care about this stuff? I'm like, I don't know, but it's so amazing, isn't it? It really is this really rich, deep, complex domain.

At the time, I remember a guy explaining to me, he was trying to explain a concept that has to do with how the body translates DNA into individual genes. This is called translation. He was explaining this, and he kept getting into more and more detail. At a certain point, I just stopped him and said, Dan, that cannot possibly be the way it works. It's just too complicated. He smiled at me, he was really nice. He was like, Chris, it's not optimising for you to be able to understand it. It's this way because it makes more. That's the only reason this works. That was scary to me. It was like, oh, this is turtles all the way down. I might never understand this.

He was like, Chris, it's not optimising for you to be able to understand it. It's this way because it makes more. That's the only reason this works.

I've spent the last 17 years trying to pick up all the molecular biology that I didn't start out with. I think that you also, in your question, hinted at a problem that I have in a bioinformatics core inside of a university, inside of an academic research institution, which is that researchers don't have any money. They can't really afford to pay you a lot of money to do the analysis. But there are these circling biotechs and pharmaceuticals all around me that are looking at my staff and saying, wow, we could really use that expertise.

So that's scary. I have to do a lot of jazz hands with my staff and just be like, hey, look over here, isn't it fun to work here? I'm mostly successful with that. But it's a tricky problem because, again, a lot of the stuff that they're doing here, they could be doing in a biotech or a big pharma, and that has a different benefits profile. I won't say anything more about that. But does that actually answer your question, Noor?

Kind of, kind of. You wanted the front door. Send an email here.

No, I'll say one more thing then, which is, you're already a bioinformatician, though, so you know this. But those three circles of biology and statistics and computer science, wherever you are, keep building those out. To feel really comfortable in that center, you're going to have to lean into these other spots at a moment's notice.

Balancing continued education across three fields

I think that's good advice, and I think this leans right into another question in Slido that's anonymous, and it says, since bioinformatics is the intersection of three highly complex fields, how do you balance continued education and professional development with all three? So if the suggestion is like, keep going with all three, get to that center of the overlapping Venn diagram that looks like a little pie piece, do you have any tips on how you do that? How do you stay in that zone? Although, I think one of the answers is that you have spent over a decade doing that.

Yes, I have. I think that is also a really good question, because one of the kind of false aspects of this metaphor that I just drew with my hands with the three overlapping circles is that those circles are actually the center of the overlapping Venn diagram. The way I draw them and the way you imagine them, they're very static, and they're very kind of regular shapes. In real life, they are not regular shapes. They are like this bubbly, effervescent mess. Balancing in the center, it's not a flat surface. You're always having to maintain the balance in the center of those three. I will tell you that in the last couple of years, we have had a couple of interesting shocks to the system. I'll just tell you one.

To do this kind of analysis at scale, an important thing that we have to do is we have to put all the analyses in some kind of an automated workflow automation system, like Snakemake or Nextflow. If you're not familiar with those, those are very cool tools. We started with Snakemake, and then the Nextflow kind of community of developers, they kind of eclipsed the offerings in Snakemake. We said, okay, well, we kind of have to transition to Snakemake, or sorry, we have to transition to Nextflow. That meant we had to learn a whole new workflow automation system, which, I don't know, the first one's the hardest, but that was a challenge.

I think you're right in pointing out that this is a dynamic balance that you have to attend. If you want to master something, I would recommend ancient Latin or ancient Greek. Bioinformatics is not that. You're always one meeting away from looking kind of ignorant about something. That's the trade-off. It's like, do you want to look smart, or do you want to learn something? If you're okay sacrificing the looking smart part, you can learn a ton at every consultation that we have. That's what's happening. I'll also say that the diversity of the projects that we get are constantly kind of testing where we should go next. Sometimes it's like, oh, well, like right now, I don't want to get into the details, because it's going to be, I mean, you'll like it, but we support a lot of different sequencing platforms. The sequencing platforms themselves keep changing. There's a new sequencing platform that we're

standing up right now called Ultima Genomics. The problem is that they generate such large volumes of data that we have to use our GPU high-performance clusters.

We're so on the edge of what they're able to do that the techs at Ultima Genomics cannot fully answer the questions that we're putting to them. Now we've got to learn how GPUs are being allocated on this high-performance cluster. Anyway, yeah, it is a constant balance across what you're learning. Fortunately, I live in a domain where the consultations are always pushing me to the edge all the time. Yeah, that's the best I could do, I think, for that. It's a great question. Thank you. Yeah, it is a great question.

The talk about Venn diagrams always reminds me of Drew Conway's diagram of data science, which is another thing that I feel like nobody's ever going to fully achieve. I'll put that in the chat if you haven't seen it. It's computer science, math and statistics, and domain expertise. It's so hard to live in the very, very middle of that Venn diagram. I don't know if it's worth striving to fully get to the center of either of these Venn diagrams because you will never, ever know everything. Just like Chris just said, there will always be something new.

Even if you lived in that very, very center of the data science Venn diagram two years ago, then AI came on the scene and you had instantly more to learn that you never would have really had to learn before. Just be nice to yourself. Okay, so fantastic question, anonymous asker. Thank you so much for asking. We had one from Natasha Lang, which is fantastic. It has a little asterisk, so I'll ask it for you, Natasha. It says, what does a typical day look like for you, Chris? Typical week?

A typical week in bioinformatics

Yeah, typical week. Typical day is, there's no such thing as a typical day for me. I actually am the director of the core, so I don't get to do nearly as much analysis as I would love to do. The people working at the core are actually doing a lot of the analyses. I'm going to create a fictional entity that's a blend of me and the analysts, but in a typical day, or a typical week, I'll say, we're going to be meeting with biologists who are coming to us basically with a research question and saying, hey, we're interested in understanding how genes are expressed when you have retinal diabetes. How are the genes different, retinal diabetes?

Are those genes different in a way where we could actually start to use that information to recommend an intervention? You should have VEGF therapy, or you shouldn't have VEGF therapy. They'll articulate that question, and it's nice when they come to us before they've submitted the samples, because we can guide them and say, oh, you're going to need more samples, or you're going to need this kind of sequencing. But sometimes, they just show up with the data in their pocket, and they say, we thought we were going to get an answer, and what we got was a bunch of data. Then we'll create a project plan, which actually, to a lot of researchers, is kind of an alien thing, like the idea of a project plan that articulates an analysis.

They thought we were just going to go off and do it while they were on the call, which doesn't typically happen. Then another huge part of the week is running the code, building the pipelines, kind of shepherding the data as it moves through the different compute systems that we have, and then trying to interpret the results.

Interpreting the results is actually difficult, in a way, because first off, the data are complicated, the tools are complicated, the visualizations can be complicated, and also, under all of this, the researchers are coming to us with novel ideas. They're running things that generally haven't been done before, so it can sometimes be a little bit elusive. You can't just be like, oh, that looks like this other dataset, because there is no other precedent for that particular analysis. That said, a lot of what we're looking at, we kind of understand the informatics, we understand how the genes work, we understand how the sequencing platforms work, so we can kind of look at it with a cup of coffee in our hand and say, that looks about right. I think that what we're seeing here is important.

But then, I think one of the most important meetings that we have is when we come back together with the researcher, and we kind of go through this and say, we think we're seeing this, but what are you seeing? That's a really tricky part of this kind of engagement model, because we have the expertise in the informatics, but they have the expertise in the domain. They really understand the disease, or they understand the biology in a way that we won't. So, kind of being able to lean into that and try to pull out the actual knowledge from these visualizations and that information, that's a really important part of our week. And also, we have to archive a lot, because there's a lot of data that's generated all the time, so we're moving data around. Yeah, so that's, I don't know, I'm going to say that's a typical week. Yeah, okay, we'll go with that.

I think it's hard to define, it's hard to fit, you know, a week or a day in the life into a description, but it's so helpful for context, so I appreciate trying to squeeze it all in there.

Hiring and the job market

And thank you, Natasha, for asking that question. There is another one in the Slido, which has risen to the top from Abigail. Again, I see an asterisk, I'm going to read this for Abigail. It says, how weird is hiring right now? Are you swamped with applications? Are some of them fake people? Are you changing anything about how you do it? Hiring is one of those questions where some people do a lot of hiring and some people don't. What experience have you had lately? Yeah, the truth is, I'm kind of insulated from that question simply because the research climate is so constrained by funding right now. Yeah, and you know, I'm looking around because there's other groups like mine on campus. We're kind of looking around and we're like, are you okay? Is everybody okay? And we're like, yeah, we're okay, as long as nothing else happens, we're okay. But I don't think anybody's really hiring actively in the academic space right now because everything is so tentative.

It does seem like there's more hiring in some of the more traditional software engineering roles. It feels like there's a little bit of a turn in the path. I think that there was a sentiment that like, oh, well, we're all going to be replaced by AI. And I think that that sentiment is shifting a little bit. That always struck me as a little bit simplistic. And I think that that sentiment is shifting. And there's, I think, an emerging idea that, oh, people are actually really important. And that AI is actually a useful tool in the hands of good people. So let's make sure that we keep engaging good people.

But yeah, I have not seen fake people, but I also can't prove that the people I've been interacting with are actually real. So I don't have a good ground truth there.

No, but you gave us a fantastic peek into what things look like in academia and in research right now. So I really appreciate that. I almost hesitated to ask that question because like, I don't think that Chris does a ton of hiring, but I know that there's good conversation to be had behind this. So thank you, Abigail, for asking that one. And this leads us right into another two questions that we have in Slido. So we're going to smush them together. One was from Noor and one was anonymous. And they both are asking about that sort of claim that AI is about to replace bioinformatics analyses or processes. What do you think about that?

AI's role in bioinformatics

It's an interesting question. I will tell you that at this moment, well, first off, I just want to back up and say like AI as a phenomenon is definitely a disruptive innovation. And it's also very clearly disruptive at scale, in a way that people are like, oh, this is like the Industrial Revolution, or this is like the invention of the printing press, or this is like the invention of a circular saw. I haven't found the perfect analogy for this yet, but bioinformatics is definitely disrupted. I think that one of the problems, and just by the way, just so that you understand the filter through which all of this is being said, I just recently realised that I am what's called a sceptical optimist about AI. I think that AI is important and it can do good things, but I'm not taking anybody's word for it. I am watching what it's doing, and I am testing what it's doing, and I am drawing conclusions as I go. So, sceptical optimist, if you want to sign up for that camp.

It's easier for me to get to this through the software engineering perspective, because people are using AI for software creation in a couple different ways. First off, they're using AI basically as what Google should have been doing 20 years ago. It's basically giving you a better answer that's articulated more clearly. So, that's cool. That's interesting.

Then, I think further down the continuum, they're using agentic agents to write blocks of code, or passages of analysis. The concern I have about this is that it's nice when you have code that works. It's even better when you understand the code that works. I feel like for many years, we were really trained in reading code and the importance of reading code as a mechanism of understanding it. Now, I think that there's a class of people that are generating code and bragging about never reading it. We just shipped this, and nobody ever looked at it. Isn't that cool? Software engineering actually has a word or a phrase for code that works that nobody understands. They just call that legacy code. That's the code that somebody else wrote that you inherited, and now you have to maintain it. Now, we figured out a way to make more legacy code faster. I didn't know that was the problem, but that's interesting.

Software engineering actually has a word or a phrase for code that works that nobody understands. They just call that legacy code. That's the code that somebody else wrote that you inherited, and now you have to maintain it. Now, we figured out a way to make more legacy code faster.

All that said, I do think the stronger swimmers are going to be able to surf this wave. I think it's actually going to be pretty cool. I think that there are going to be weak swimmers that are towed out into dangerous currents. I'm a little bit worried for them. We try not to do that when we're engaging researchers, but researchers also come to us and are like, hey, I ran my data through this AI. What does it say? We're like, okay, well, let's back up. What was your prompt? Let's look at the code together. I've had people take code bases and run it through AI to refactor it and basically make it easier to understand and maintain. Then come back to me and say, well, I understood the old version, but I don't understand the new version. Can you explain why it did this?

I don't know. I think that research is really important. I think that computationally reproducible research is really important. I think an essential ingredient in that is understanding. Frankly, I think we're still on the learning curve to understand how can we deploy AI in a way where we're not compromising our human understanding and the patterns of collaboration that can really advance the domain and still get the benefits of a large language model on that existing knowledge. It's going to be cool in five or ten years to look back on this and realize we've all been using Roman numerals the whole time. This is so much easier now.

I wish I had the answer. If I did, I would tell you, but I don't. I just feel like it's a little early to figure that out. We're all still figuring things out, but it's wonderful to hear varied opinions and experiences from different people. Thank you for sharing yours. Sometimes, I think that we just cycle through the same problems in different ways. I'm looking at the chat. I know that you're not looking at the chat because I don't want you to because it's a lot. There's a conversation in the chat that's like getting access to high-end GPU is a bottleneck for us right now. It makes me think about all the conversations that I heard growing up, which were trying to get time on the mainframe at campus and using tape and stuff. It's the same problem, but it's different because now the capacity is just so much higher. We are still in the same cycle. It just looks different. It's a different flavor. If no one's experienced that or heard stories of that, like I was born in the 80s when people were trying to get time on the mainframe at campus. I didn't experience it, but I heard lots of stories.

Specialist versus generalist

Thank you for asking that question, Noor, and our anonymous asker. I really appreciate it.

I saw a question from Zach that has an asterisk. I'm going to read it. It says, what are your thoughts on specializing in one specific domain versus being a generalist? I really like this question. It's interesting. To a degree, I think that's almost a personal decision. I'll also just point out that I had no foreknowledge of how I got to here. Nothing

makes sense to me in the moment. Why are you doing this? Everything makes sense in retrospect. I just feel like I've been in this beautiful canoe ride down this beautiful river and facing the wrong way the whole time. I'm a little hesitant to give you any advice on how to steer, because again, I'm generally facing the wrong way.

For me, I really like the interdisciplinary aspect of bioinformatics. I feel like I'm often at the interface of a couple fields, and also at the interface of people, in the sense that there's the people at the high-performance compute environment here that are like, oh, well, we have these kind of GPUs, but we don't have those kind of GPUs. We've got these GPUs, but you can only queue them up this way. Almost nobody really understands when they're talking, and I can step in and be like, okay, I think what they're trying to say is, which is probably a rude thing to say if they're actually on the call, but that interface, that translational aspect of, I understand the informatics, you understand the molecular biology, and together, we can have a super interesting conversation.

To get to your question, for me, I felt like I really had a deep background of computer science and engineering, so I had been scrambling to learn the statistics and to learn the molecular biology. and I found that those things just kind of had their own wellspring of enthusiasm. So I was just chasing, I was chasing what felt important and felt good at the time.

I think that if you look at, I mean, if you look at the Drew Conway, did you put that on the Slack, Libby? I did, I put it on the, in the chat. Okay. The Venn diagram, yeah. Yeah, yeah. So that's perfect. So I mean, like, if you look at that, and there's a couple variations on that, but I think having exposure to the different petals of that Venn diagram, I think it's super useful because it gives you the flexibility to have those different conversations and to kind of, like, move among those different activities. Where you actually land and where your sweet spot is, this is, I think, a little bit more personal. The flip side of that is, though, I would definitely get enough exposure in some of those other domains to feel comfortable. Like, and I guess, to say it a different way, there are times where I'm like, ooh, I don't feel like I really understand the linear mix model well enough to explain it to this PI. So that to me is a sign, I need to get over there. I need to, I need to go deeper on this because I want to be able to explain this to the PI, sorry, the principal investigator. That's, that's, that's kind of our customers. So I want to be able to explain to the customer. So I got to go a little bit deeper over here so that I can actually have that conversation. So for me, it's like, it's almost this weird thing of like, I'm afraid of this. I should go closer to it.

Um, so, but I mean, ultimately, like, I think it is up to you and like where your sweet spot is in that domain. I think if you go deeper, you're going to have that experience and, and you're gonna have that depth. But if, uh, if you stay, if you stay a little bit, a little bit shallower, then you're going to have that breadth. I like breadth. That's just where I'm at.

Yeah. I will bring up again, the book range, go check out the book range. Um, I'm a generalist by nature, just because I'm easily bored and I, my brain likes to hop around and do everything. So I will also recommend the book refuse to choose by Barbara share all of Barbara shares books. Really? If you're a person like me, who she, she calls scanners, um, or multi potentialites. I think that there is, there are a lot of people who don't choose actively choose that as a career path, being a generalist, but they kind of are by nature. There's a reply to Zach's question in Slido that I wanted to just read before I mark it as answered. And it says, if people stay in a role long enough, they will pick up domain knowledge. I prioritize coding skills, being a good fit for the team and the ability to ask good questions, which is what helps you pick up domain knowledge. Yes. Fantastic. Whoever you are anonymous person who, who answered this is perfect.

Yeah. Okay. Sorry, if I could just piggyback on that, if you are, if you were, if you were in a place in, in kind of your own personal equilibrium where you can, um, where you're comfortable asking a question where you're comfortable, basically, and I'll put a finer point on it. If you're comfortable asking a dumb question, you will open up such an interesting conversation. And it's not just a conversation where you learn more. It's a conversation where the whole collaboration opens up, like, because I will tell you that domain experts come to me and I do not know what they're talking about. And I'll just say like, can you put a few more words around this thing that you said? And, and they love to talk about it. And so they're like, absolutely. I'm so glad you asked that question. Um, so I think that the willingness or the ability to be a little bit vulnerable just creates such a more satisfying and deeper and productive collaboration. So if you can do it, it's, it's a huge win.

If you're comfortable asking a dumb question, you will open up such an interesting conversation. And it's not just a conversation where you learn more. It's a conversation where the whole collaboration opens up.

AI, human collaboration, and trust

Wonderful. Thank you to both of you. Okay. Um, I have a question from Nellie. Nellie, our, our resident treadmiller these days. Um, I will put your question in the chat right now. Wonderful. Uh, I wanted to, or hi, my name is Nellie. Thank you so much for joining us today. Uh, I am curious. Um, a lot of people tend to ask like how people are using AI or how that's affecting their life or their work and their lives or whatever. I'm curious, what human aspects are you leaning into, especially as you mentioned, you're skeptical optimist. And I kind of agree with that phrasing. I'm also kind of in that space. And so I was just kind of curious where you lean into the human aspects of work and other things like that.

Yeah. I mean, so it is, it's really interesting question. Like, like in the age of AI, what does it mean to be a good human? Um, uh, and I mean, in my case, what does it mean to be a good researcher or what does it mean? How can you support good research? Um, AI is a, is an interesting tool in that. Like you can come to AI with a research question and you can just put it to AI and AI.

Well, I'm going to share a little conversation that I had with Claude, the agent, like not too long ago, like I was in, I, I, I sat down, I was in a meeting and the meeting wasn't going well. And so I started a conversation with Claude and I said to Claude, um, how can you achieve computationally reproducible research results when you're using a nondeterministic tool like AI? And it said, well, it's harder, but there's ways to do it, but that's not really the problem. The problem is that people over-claim like the impact or the significance of their results. And I said, and I kept going and I said, well, okay, but in a domain that rewards, um, positive results over negative results, how is using a tool to do that? And he said, well, how is using a tool that is optimized to offer really confident, comprehensive answers? How is that going to shape the direction of the research and the science? I mean, won't that just reinforce positive over-claiming? And it said, this was the most disarming answer I've ever gotten from an AI. It said, yeah, that's a really good question. This might get worse before it gets better, which I was like, that's such a gangster answer. I was totally unprepared for Claude to be that like vulnerable.

Um, but it also was really scary. Um, I think one of the problems that I have with AI as a kind of faux collaborator is simply that it isn't a collaborator, that it is not teaching us to be vulnerable. It is not teaching us to be better. It's not, it's not helping us to trust other humans more. Um, and so I feel like it's kind of, um, encroaching on the nature of collaborative research, um, which I think is built on trust. And I think trust is built on vulnerability. And I, I never, I like almost, almost never, except for that one moment that I just described, Claude doesn't say, I don't know, like it's hard. I mean, it does, but you have to push it really hard. Um, and so to, to interact with people that are like, I don't know, I don't know the answer to that. I'm unfamiliar with this. That's a whole different muscle that AI doesn't exercise. And I think it's really important for research.

I think one of the problems that I have with AI as a kind of faux collaborator is simply that it isn't a collaborator, that it is not teaching us to be vulnerable. Um, and so I feel like it's kind of, um, encroaching on the nature of collaborative research, um, which I think is built on trust. And I think trust is built on vulnerability.

Um, so I, and another thing that, uh, that Libby and I talked about a little bit, um, earlier, uh, off offline was this notion, uh, it's called Conway's law, Conway's law. It's by this guy, Melvin Conway, who was a computer scientist, uh, in the mid 1900s and Conway, he observed this thing where it's like, uh, basically what he said was all other things being equal, the architecture of a technical solution will mirror the organizational structure that created it. And specifically what he said was the architecture of the solution is going to look like the communication structure of the organization that created it. So if you have like three chief architects and you say, you guys got to go build a compiler, they're going to build a three pass compiler, right? Because there's three chief architects, that pattern of like code reflects the thing that created it, right. That can be bad and that can be good. Um, but I do wonder what relying on AI is doing to that pattern. I'm wondering like, is AI creating patterns that are easy for us to consume that fit into the organizations that are, um, easy for us to understand and communicate to other people answers? I'm not sure. I'm just not sure yet.

So, um, I think if I can just the short summary of that is I rely on people to be people and I rely on people to remind me how to be vulnerable and to build trust and like to collaborate. Um, and you know, and those people are using AI and I don't begrudge them that. Um, but I'm not getting, I'm not, I'm not exercising that muscle very much with AI. So I rely on people for that. Does that, does that get at it? I think it gets at it enough. I think, I don't know right now is a valid answer to pretty much anything. It's always a valid answer.

Moving from academia to industry

Um, okay. I have another anonymous question in Slido here. Um, which is I recently complete, I recently completed a highly statistical masters in statistics and data science bioinformatics in Belgium, but I realized that academia and interest in industry are quite different. Can't talk today. What advice would you give for successfully moving into industry as a recent graduate?

Any thoughts here on moving into industry and academia and industry are, that's a good question. I mean, I actually came from biotech that was, um, I mean, yeah, it was commercial biotech and I went to academia. So that was a different, that was, that was a different route. Um, well, it was confusing. There is, I mean, like the academic route, you know, is optimized for knowledge generation and that creates some interesting patterns and kind of in the code basis that you end up using. Um, they tend to be a little bit more exploratory. They tend to work really well in the labs that designed them and don't always generalize. Um, commercial solutions are typically optimized, um, along different axes. I mean, they, they, they have to be robust. They have to be, uh, scalable. Um, they don't necessarily need to generalize across a whole bunch of different users. Um, so it's different.

Um, I think it depends. I'm sorry. This is like, sorry, as a bioinformatician, every conversation that I'm going to have with you starts off with the phrase, it depends. Classic statistics answer. Okay, good. So yeah, there you go. You know what I'm talking about? Um, I think it does depend a little bit on like, I guess what I'm, I'm thinking about, like, I'm not really sure that I would treat it too monolithically in the sense that like industry is not industry. It's like, it's like, are you going into a small biotech that's trying to like try to find biomarkers? Are you trying to, you know, go into, um, like, uh, a sequencing platform that's trying to build out visualization tools or analytical tools? Um, I think it depends a little bit on like what direction that you want to go.

Um, I will say that you can get an immense amount of depth in the industry side. Um, but I think if you want breadth, you can get more in the academic side, um, especially in a service kind of slot, like where we are, where we're dealing with like, um, you know, 75 to a hundred PIs every year. Um, so they're always coming to us with something new. Uh, we don't give as much depth though. Um, so I think it depends. I think it depends on the organization that you're looking at or the kind of sub domain that you're looking at. It definitely does. Gonzalo in the chat says industry be prepared for messy datasets. Yes, absolutely. And if you would like to learn more from somebody who is at a small biotech place, um, I really recommend watching the data science hangout with Varun Dwaraka. He is the director of bioinformatics at True Diagnostic. Um, speaking of looking for biomarkers, I will put the YouTube video URL in the chat. Oh, I should have like put a little thing there that says this is Varun here. I'll put a little note there. Um, that might give you some perspective, whoever our asker is. Thank you for asking.

Okay. So we have just nine minutes left. I would love it if we could sneak in a question that we always try to sneak in, which is a piece of career advice. Is there something you wish that you knew before you got into the domain you're in or a piece of advice that you like to give others that always helps? So weird. It's,

it's, this is a weird answer, but I think, I think a lot of people on the call might get it. Um, frankly, I wish I knew more linear algebra before I got into this. Um, um, and I can't say that in a lot of venues and have anybody understand why or like what even that means, but, um, I think that for me, I, and actually the interesting thing is that like, I, again, I came into it with computer science and software engineering. So that was the, I had already checked that box, but, um, the molecular biology was almost self-powering. It's like, I'm really interested in this. So I got to learn about this. This is fascinating. Um, the statistics was the hardest thing for me, like in learning the statistics and learning the statistics well enough to be able to explain it to somebody that doesn't understand statistics and maybe even doesn't quite believe statistics.

Um, so yeah, I think that, uh, the, the, the basic career advice, if you want to go into bioinformatics, you're going to want to take a tour of duty in like all three of those domains and just suss out, like, do I like this? Is this, and if you do, then I think that they're like staying in the middle, it's hard, but it can be really rewarding. But if you get in there and you're like, I don't really like molecular biology, um, you still have a vibrant data science career ahead of you. If you, if you get in there and you know, you find out that you don't like computer science, like there's still like that Venn diagram, those three overlapping circles, all of the pedals have like, like, those are, those are vibrant and powerful careers. So, um, but if, yeah, if all three are lighting up, then yeah, bioinformatics could be a really cool thing. Um, yeah, I don't know. That's what I'll say.

staying in the middle, it's hard, but it can be really rewarding.

That is good enough. Well, I tell you, you have people agreeing with you in the chat and I will agree with you a plus one for linear algebra, as far as like, if you would like to up your game in anything and linear algebra has been left out of your curriculum, the way that it was left out of my two degrees that we're all about statistics. Um, it under underpins so much of all of the things that you might otherwise learn without it, uh, like PCA, like all kinds of things. So if you're, if you're like wondering, well, what is a vector? What is an eigenvalue? What, what are all these things? If you even scratch the surface of linear algebra, I think that they will all make more sense to you.

Data curation and working with messy datasets

Okay. Six minutes left. Can we squeeze in more questions? I think that we can do it. Let's see. I have a question from Jared Allen here. It's got asterisks and I'll read it. It says working with academics, do you find that you have to commit more time than you would like to data curation and data wrangling, or do you insist on then submitting curated datasets to you?

Yeah, good. Great question. Like some, somebody said earlier, get, get, you know, if you're going into industry, get ready to work with, uh, with some kind of hairy datasets. Yeah. Hairy datasets is probably a good starting position for everybody. Um, like we're dealing with a lot of sequence data. Sometimes it's proteomics data, but mostly it's sequence data. And we collaborate with a group that runs a bunch of sequencers, um, that, uh, that literally work right behind me. Um, so the datasets that we get are often, are often fairly rectangular datasets. Um, and they're annotated well enough for us to run the experiment, but it is also the case that people come to us and they say, Hey, I want to run this, but I also want to use this public dataset that's over here in this public repository.

Now it's very cool that there are public repositories for all of our, all of, uh, or almost all of our datasets. That's powerful. That makes our research so much more impactful. The problem is the public datasets are not evenly curated. They do not have controlled vocabularies yet about how the samples are characterized, um, and described. They are not necessarily, uh, normalized along the kind of molecular measurements that are, that are being made on that platform. So, and, and the problem is I love the idea. I mean, I got to tell you, it's kind of a novel idea, uh, of just saying like, Hey, no, you got to clean this up before we look at it. The people that are coming to us, they have no idea there's anything wrong with these datasets. They have no idea like, like how much effort it's going to take to make this into a clean rectangle that we can actually put into any kind of a model. Um, and even if they did, they wouldn't know how to do it. They wouldn't know how to clean it. Um, so that is a lift. We kind of absorb that. We will tell them that and say like, just so you know, what you're asking is hard for surprising reasons. Um, and, and they assume that it's like, Oh, it's in a public repository. It must be super clean and awesome and a perfect rectangle. And it is generally not.

Um, so yeah, um, I, I, I, I, I'll put that with the teams. Like, could we just say, no, can we start saying, no, I think I would be, I've been thinking about this as you've been talking. I think if somebody else cleans the data and then gave it to me, I would be missing so much context in that missing, missing data. I would be missing things that would have allowed me to ask questions about it and learn more about the dataset. So I think that the person who's analyzing it, getting that mess is actually maybe the best thing. I don't know. What do you think?

AI's role in cleaning electronic medical records

Yeah. I will say this, that like, sometimes we work with electronic medical records and this is one case where I almost feel like I might have a, I might have a little love for AI because electronic medical records are very informative and they are impossible to work with. Um, so the idea of AI kind of distilling that into fielded computable information, it's not that that's not a crazy idea. Like now there's still hallucination and slop. Um, so you have to trust and verify, but yeah, I could, I could, I could see that as helping. And in fairness, like our, like the AI tools, um, do actually a reasonable job of taking weird kind of, um, nominally incompatible datasets and cleaning them up and making them kind of, you know, making them, you know, stackable and, uh, and regular. Um, so it does make that easier, but you really have to pay attention because it, it, it doesn't always make, it doesn't always make perfect decisions. And if you go fast, you could, you could do something that would be, um, incorrect. Yeah. Yeah. Takes that old, uh, move fast, break things. It turns into a warning, which is like, uh, if you move fast, you might break things.

Missing software engineering

Okay. I have a question that I would love to stick in our last one minute here. It's from Elizabeth and it has an asterisk. It says, do you miss anything from the strictly software engineering side? Uh, I really, yeah. Do I miss anything from the strictly software engineering side? Um, well, my first answer is like, I miss the clarity of it because again, it really is turtles all the way down. Like you, you, you don't always have a ground truth. But before I worked in the biotech, I worked, uh, in a company that, um, that built software to process dental claims. And I will tell you that having done that for just seven months, there is emerging complexity in processing dental claims. Like you think that like somebody somewhere figured this out. I'm telling you that that's just about as mysterious as molecular biology.

Um, so yeah, you're giving me new appreciation for my mom who used to be an insurance specialist at a dental office for a long time. I've never known what she does. Maybe I should start asking questions. Don't start something you don't want to finish, but yeah. Uh, yeah, yeah. So yeah. Um, yeah, I don't know. That's a great question. What do I miss? I don't have a really good answer for it. Um, Chris, Chris will be thinking about this later and post about it on LinkedIn tomorrow or something. Well, thank you, Elizabeth, for asking that question. You have a great name. It's my name too. Uh, this was really, really fun. We've reached the top of our hours, so we have to say goodbye. Thank you so much, Chris. I hope you had a good time. I had a great time. Thanks everybody.

Wonderful. You can find Chris on LinkedIn, Chris Gates, University of Michigan, and you can also find all of us after the call on Discord. So please get on Discord. We would love to hang out with you there. Thank you for hanging out with us, everybody. If there is good stuff in the chat that you don't want to lose, like someone's name you were having a great conversation with, save it, three dots in the top, save chat, find them on the Discord, post on the, the, uh, data science hangout channel on Discord and be like, who is this person who we were having a great conversation? Um, so we can connect you to each other. Libby, can you say and just test that out real quick? Cause somebody told me they weren't able to save the chat and I just want to make sure if other people can't, can you let me know? Yeah, let us know. Somebody click the top three dots right now, unless you're in the browser in zoom, which often is different and you might have three dots in the bottom right of your screen and let us know. Can't save chat on mobile. Okay.

Okay. Isabella put it in the chat too. So can people download that? I'm just trying to figure out if it's when you're on certain browsers or for certain company firewalls too. Yeah. Cameron says can't save it on the web browser either. Um, zoom makes changes all the time. They don't tell us, they don't ask us if we just, you know, deal with them as they come. Um, but yeah, if you do not see it, I might also be able to just pop it on discord. It's not a reason to get on discord. I can download it, um, after the fact. Okay. Steve says it works on safari. Oh yeah. And it goes into a very weird spot. Um, so when I, when you download it in zoom, it shows like, can you show this in your finder? It should show that. And then you can find like a zoom folder and hopefully we cut this out from the recording, but I just wanted to figure that out with everybody. Okay. Everybody. Yeah. And if you have more updates on things that did or did not work, uh, go to the discord, Jeffrey, I will see you on the discord. I'll put the chat there for you.

Thank you everybody for hanging out with us. If you are going to join us on Tuesday, and I highly recommend you do at the data science lab, Nick crane's going to be joining us. They're going to teach us all about parquet files. I have so much parquet FOMO. You have no idea. And parquet is so valuable and powerful. Chris is giving a thumbs up. Is this an endorsement for parquet file? You need it. You need it. Okay. So come learn about it with Nick and with me and with Isabella at the data science lab. It's a messy screen sharing place where we just hang out and have fun. We are totally in the discord for chat. So it's another reason to get on the discord and get acclimated there. We're in the data science lab channel. And then if you are joining us on next Thursday, we're going to have a very special guest, Amy Rossi. She's the chief people officer at Posit. So if you've ever wondered about people analytics, about, you know, what decisions, what data helps decision-making in people analytics in the people area in general, she's wonderful. And it's going to be a very good time. Thanks for hanging out with us, everybody. See you next week. Bye.