3 | Arjun Raj’s bag of tricks

In this episode, Itai and Martin talk with Arjun Raj, Professor of Genetics at the University of Pennsylvania. Arjun understands the functioning of biological cells using a bag of tricks that he carries from problem to problem; the art of science, he posits, lies in figuring out what tricks will tell you what answers to what problems. Arjun thinks that we are all born night scientists, and that it's day science that needs to be learned. The ultimate goal of life as a scientist, he believes, is not so much writing papers, but building people. Arjun has received many prizes, including the NIH Director's New Innovator Award. His lab pioneered tools for studying biological processes using state of the art imaging and sequencing technologies. 

For more information on Night Science, visit www.night-science.org .

Transcript

Arjun: It's like, you know, we're all, in a way, artists doing basic science, right?


Arjun: We're artists trying to, you know, carve a sculpture out of this raw block of empiricism that the world is.


Arjun: Like, we're trying to make sense of it in some way, and to give it form and shape.


Martin: And to find the statue within.


Arjun: Exactly, yeah.


Itai: Welcome to the Night Science Podcast, where we explore the untold story of the scientific creative process.


Martin: We are your hosts.


Itai: I'm Itai Yanai.


Martin: And I am Martin Lercher.


Itai: So, it's great to have Arjun Raj today.


Itai: You did your undergraduate at UC Berkeley in both math and physics.


Itai: And then you came to NYU for a PhD at the Courant Institute before taking a sort of left turn into biology.


Itai: Today, you're a professor of genetics.


Martin: I was actually surprised, Itai, that you said it was a left turn into biology.


Martin: I would have thought it was a right turn, but...


Arjun: I'm a lefty, so the left is good.


Martin: So, the research that Arjun does is broadly in RNA systems biology, looking at mRNA and non-coding RNAs and with a particular focus on cancer biology.


Martin: Arjun studies the biology of single cells mostly.


Martin: And the reason for that is that even genetically, identical cells, as Arjun keeps stressing, can show very different behaviors.


Itai: Well, you know, Arjun and I are sort of parallel in the sense that we have some C. elegans work and we have some cancer work now.


Arjun: Yeah, that's actually true.


Itai: I think I never thought about it that way, because I think of you probably the same as I think of me, which is at the heart of it, we're gene expression people.


Arjun: Yeah, I think that's kind of where I started out.


Arjun: And so everything else just feels like these little dips into interesting systems.


Itai: Arjun, we're so pleased to have you.


Arjun: Yeah, thank you so much.


Arjun: Thank you so much for having me.


Arjun: And thank you for the very succinct summary.


Arjun: Did you want to add anything to that?


Arjun: It's always kind of humbling to be like, you know, to have your career is like, well, this is about like  seconds.


Arjun: I was actually thinking a little bit about, you know, what are our contributions?


Arjun: I was thinking like, what will be in my tombstone?


Arjun: And it will be like, made some interesting observations and some tweets.


Martin: So, you know, if you were to design your own tombstone, what would you write?


Arjun: Oh, dude, why did I open myself up?


Arjun: Actually, though, you know, I have thought a little bit about, you know, what is a legacy or, you know, a scientist product is certainly papers, like making papers is what we're in the business of doing, for better or for worse.


Arjun: But then, you know, I've become less interested in the goal and a bit more interested in the process, like how do we systematize thought?


Arjun: How do we teach things like creativity, right?


Arjun: In the end, what you're building is not even papers, it's people.


Arjun: You have a way of thinking about the world, like a school of thought.


Arjun: That is really your legacy is, you know, how many people you were able to convince to think like you do for whatever value that has.


Arjun: And in the end, I mean, it's all an ecosystem.


Arjun: There's many different ways of thinking about different things.


Arjun: But I think that would be, you know, perhaps not an epitaph, but a living epitaph that I would be proud of.


Martin: Okay, so your tombstone would say he left a school.


Arjun: Yeah, I suppose so.


Itai: School of Arjun, if you're propagating a school, I'm wondering, you know, did someone have you in their school?


Itai: Where did you learn how to think?


Arjun: I grew up always wanting to be a scientist, but I always really wanted to be a theoretical physicist.


Arjun: I went to undergrad at Berkeley and I got really interested in, you know, I started taking a bunch of physics classes.


Arjun: When I met an advisory, he said, well, you know, if you really want to do theoretical physics, you should do math as a double major.


Arjun: So I said, okay.


Arjun: I started taking some math classes and I really liked it.


Arjun: It was so beautiful and abstract.


Arjun: At some point I realized I just kind of wasn't like smart enough to be a mathematician.


Arjun: So then I thought to myself, well, you know, some sort of applied math might be a good way to go.


Arjun: And I applied to the Courant Institute, which is, you know, really known for applied math.


Arjun: And, you know, I got there.


Arjun: I didn't quite know what I was thinking about doing, but so I started working with this professor there, my advisor, Charlie Peskin, a truly wonderful person and brilliant, brilliant scientist and mathematician who also, by the way, had a really interesting path.


Arjun: I mean, he was in medical school and got into math by trying to model all kinds of fluid flow.


Arjun: And he's open to interesting paths.


Arjun: I remember, okay, I'm into bio-math.


Arjun: I should say that I was like, I'm never working in a lab because like my worst grade in college was in lab, the physics lab, where I was like, I hate lab.


Arjun: But I thought to myself, if I'm going to get into molecular biology, again, purely from a theorist perspective, it would be kind of good to work in a lab.


Arjun: So I thought, okay, the summer after my first year in grad school, I asked Charlie, I said, hey, do you know a lab I could work on?


Arjun: He's like, yeah, my son has this friend and I play tennis with his dad.


Arjun: He runs this lab out.


Arjun: So that's how I got introduced to someone named Fred Kramer, a really kind and brilliant scientist who actually, by the way, also invented Sanger sequencing, but he did it for RNA instead of DNA.


Arjun: And so it took him longer to publish.


Arjun: Anyway, super smart, brilliant guy.


Arjun: And Fred, he kind of said, hey, why don't you work with a sub PI that I have in my lab, this guy, Sanjay Tyagi.


Arjun: So I started working with Sanjay that summer.


Arjun: And I, for whatever reason, I just loved it.


Arjun: I fell in love with lab work and I said, you know, I would love to do this for my PhD.


Arjun: And they were so kind and open.


Arjun: And, you know, Sanjay ended up being my co-advisor and I did most of my work out at this lab.


Arjun: So I had probably the single weirdest PhD from Courant.


Martin: That seems very unusual for a mathematician.


Arjun: It is highly unusual for a mathematician.


Arjun: It was a glorious time though, working with Sanjay.


Arjun: It was a, you know, a small group and, you know, I'd have lunch with Sanjay every day.


Arjun: And I feel like, you know, getting back to this point of learning the creative process, I learned so much of that from Sanjay.


Arjun: He was just such a creative out of the box thinker.


Arjun: Like, you know, he had ideas.


Arjun: He would do a lot of experiments himself.


Arjun: And I remember he had, you know, he was pitching some idea to me.


Arjun: It's like, I even did the preliminary work to show this thing works.


Arjun: And then like literally  years later, I see a famous lab published it in Science.


Arjun: He just had like all these cool ideas, really pretty visionary and creative.


Itai: Did he have a crack in the wall for you to peer through his process?


Itai: Or would he sort of just tell you about the results?


Arjun: I don't know about a crack in the wall, but he had an open door to his office.


Arjun: A lot of it is we would just, I would just hang out in his office and we'd talk about things and he'd just have all these ideas that were just so kind of creative and clever and brilliant.


Arjun: And in prepping for our discussion, I was trying to think like, like how does it just rub off on you?


Arjun: And on some level, the question is what is creativity?


Arjun: And the more I think about it, I think at least part of it is just how big a bag of tricks do you have?


Arjun: Like how many different tricks have you seen applied in different contexts?


Arjun: And perhaps some people have a little bit more skill at remembering all those tricks or applying them in different situations.


Arjun: But yeah, I think that what I learned from Sanjay really was maybe just a bag of tricks.


Arjun: I'm sure I don't have all his tricks, but maybe that is some version of the creative transfer.


Martin: But when you say tricks, like what do you mean?


Martin: Like experimental techniques, probably not.


Arjun: I'll give you a very clear example.


Arjun: So, in his lab is where we developed a lot of these ideas about transcriptional bursting.


Arjun: Is it the local environment of a gene that determines whether it bursts or not?


Arjun: And so Sanjay had this idea, which was, well, if we integrate two copies of a gene in different locations in the genome, which I had done it that you would see uncorrelated fluctuations, which I had seen.


Arjun: And then he said, hey, what if you take these two colors of a gene, you know, this is the old days, just do it in trans genes and co-integrate them.


Arjun: So they'll integrate into the same locus, which is always a technical kind of artifact really.


Arjun: And when it integrates in the same locus, then you should see them burst together because they're right next to each other and you can see the two colors.


Arjun: It's just such a beautiful experiment.


Arjun: And it was a very clear result in that paper.


Arjun: So that's the thing, right?


Arjun: It's like an evolution of that trick.


Arjun: These tricks, I feel like they just propagate over time.


Arjun: Actually in math, it's kind of the same way.


Arjun: There's, you know, you have this like bag of tricks and to the outsider, it looks like brilliance or quote unquote creativity.


Arjun: And I think part of it is we try and create a mystique around that.


Arjun: But if you just demystify it and just say, look, I just have this trick and I applied it here, and you can also have these tricks and apply it wherever you want.


Martin: Yeah, if you would just have one trick and would apply it to system after system, that wouldn't be very creative.


Martin: But I think part of the creativity is to understand at what point you should use which trick, right?


Martin: And so I do think that is creative.


Arjun: It's a question of, well, how many tricks do you have at hand?


Arjun: So when I started my lab, it was very much kind of one trick pony, if you will.


Arjun: Like here's this one thing that we do.


Arjun: And there's a clever art to figuring out what are the problems you can answer using that trick.


Arjun: And that's certainly a fine way of going about things.


Arjun: I think now that the lab has gotten a little bigger and we have more expertise, we're able to take a problem and ask the question, what tricks can we apply to solve that problem?


Arjun: And there is an art to figuring out what tricks will tell you what answers to what problems.


Arjun: And I suppose that's still a bit of an art, but you could reduce that to saying, I have this set of tricks, I have this set of problems, and I'll just try and cross match them until something works.


Arjun: I mean, maybe that's all our brain is really doing.


Martin: And I mean, maybe that's all our brains are doing all the time.


Martin: I don't know.


Martin: But do you think that you have a finite list of tricks?


Martin: I mean, like, you know, you have , ,  tricks that you could write down, and then everything you ever did in your papers would be somewhere on that list?


Arjun: Probably, yeah.


Arjun: I think it's probably like that for almost everyone really.


Arjun: I mean, I remember someone talking to me about some, you know, very famous and creative senior scientists and saying, you know, this person, their lab's a mess.


Arjun: It's completely disorganized.


Arjun: Everything is crazy.


Arjun: But I will say that they have a really good idea about every five to seven years.


Arjun: And I was thinking, I was like, that's actually really good, you know?


Arjun: Like to have a new creative idea is pretty hard.


Arjun: It's pretty hard and it's quite rare.


Arjun: And, you know, there's nothing wrong with being a one-trick pony or the, you know, one hit wonder from the s.


Arjun: That's fine, man, a lot better in being a zero hit wonder from the s.


Itai: And also if you have a one, you're a one-trick pony, but you're creative in knowing where to apply it, like which field to go to.


Itai: That's also deserves some credit.


Itai: To me, it seems obvious from your background is that you picked up a lot of tricks by being interdisciplinary and coming in from math and physics.


Martin: Which is something I have a lot of sympathy for.


Martin: My own background is in physics as well.


Arjun: Yeah, I was thinking a lot about that in the context of night versus day science.


Arjun: And for people who are interdisciplinary, I feel like, okay, if you're interested in interdisciplinary work to begin with, you're kind of born a night scientist.


Arjun: Your instinct is to make connections that are probably wrong or not there, but that's every so often they might be, and that's wonderful.


Arjun: So I feel like, yeah, I was kind of born a night scientist and trying to find correlations, perhaps many of them dubious or spurious, between different fields.


Arjun: And a lot of my growing up as a scientist, which I feel honestly is mostly happened in the last five years or so, has been to learn day science from biology.


Arjun: How do you give substance to your art?


Arjun: And that's been kind of a profound change for me over the last five years.


Arjun: And I think we do better science because of it.


Martin: That's a very interesting perspective.


Martin: But yeah, when we were discussing who we should talk to next, and Itai suggested your name, he introduced you as somebody who might actually do more night science than day science, which is rather unusual.


Martin: So it's interesting that you kind of confirm that by saying you're slowly learning the day science part.


Arjun: Yeah, I think that's true, yeah.


Martin: But our impression is that overall, it seems that a lot of people focus much more on the day science part, and almost forget that there's this wonderful night science that we do.


Martin: I mean, of course, everyone has to do night science, but we feel that a lot of people don't give it enough space in their days.


Arjun: Well, I think of science as kind of an ecosystem.


Arjun: We have different people working on all kinds of different problems.


Arjun: Some of those problems are important.


Arjun: Some of them are interesting.


Arjun: Some of them are creative.


Arjun: And there's some Venn diagram where you can put all your work along these different axes.


Arjun: I've come to appreciate people who do more day science as a person, I think, who really started out doing night science almost exclusively.


Arjun: When a scientist is born, right?


Arjun: When a kid shows interest in science, for sure, they're not thinking...


Arjun: I hypothesize that they're not thinking day science.


Arjun: I think we're all kind of born night scientists.


Arjun: And I guess one of the big challenges is how do we learn to do day science without giving up on night science?


Martin: I think that's a very deep thought.


Arjun: I mean, in a way, it's actually very interesting that, I think these days many junior faculty, they're really worried about getting grants.


Arjun: And I totally get it, right?


Arjun: I mean, getting grants is important, you need to do it.


Arjun: But if there's anything that's more day science than getting a grant, I don't know what it is.


Arjun: And once that becomes the point of your existence, then hope is lost, yeah.


Arjun: I mean, why are we doing it then?


Arjun: I mean, surely there's easier ways to get money.


Itai: So how do you train your students then, the school that you're building, how do you get them to remember about night science?


Arjun: I don't know if it's a selection thing.


Arjun: I really feel like people are all night scientists at heart.


Arjun: That's why they got into it.


Arjun: So I feel like actually, the people in my lab, most of them or all of them really have a passion for learning something new, finding something fascinating, unexplained, seeing the gorilla in the room.


Arjun: I think all of them want to do that.


Arjun: And actually I think my training is usually to provide them A with some more tricks in their arsenal.


Arjun: The real challenge is how do you convert those ideas, like night science ideas into something you can actually do during the day.


Arjun: And then also, how do you make it really rigorous?


Arjun: To me, the biggest failure of night science is in looking at alternative explanations.


Arjun: So another way of thinking about things other than night and day science is there's work that's fun and there's work that's not fun to do.


Arjun: And then there's the work of thinking.


Arjun: So another axis is the work of thinking and the work of doing.


Arjun: So you might think about a project or you might actually do the experiments.


Arjun: And I think a lot of us equate, the fun work is thinking about the ideas and then the not fun work is the actual doing of it.


Arjun: And we could almost think of this as a night and day thing.


Arjun: So night is sort of fun thinking and day is not fun doing.


Arjun: The one thing that I feel like, frankly, is a problem in systems biology in general, and this is actually something I learned from a postdoc that I had early on in the lab, is I think we avoid a lot of work that is simultaneously not fun, but thinking work.


Arjun: And that's like, I have this cool idea to explain something.


Arjun: Now, let me sit around and think for a week or two about the boring ways to explain my data.


Arjun: Here's another way I could get this behavior, and it's kind of boring, but I need to then either exclude or include that.


Arjun: We have to then do another experiment to figure that.


Arjun: I mean, that is kind of the scientific process, and I think that's also very much day science.


Arjun: It's a thing that I had to learn.


Arjun: I can even just give you a concrete example of this.


Arjun: So we have this paper about cell size regulation sensing.


Arjun: There's this question, when a cell of a particular size replicates its DNA, now all of a sudden you have two copies of a gene instead of just one.


Arjun: So the question is, in order to maintain the concentration of the RNA, how does it produce half as much RNA from this now doubled amount of DNA upon gene replication?


Arjun: And basically I had this thought at the time that right at the time of replication, perhaps there's some mechanism there, you could imagine some sort of chromatin mark type thing, but there's some kind of mechanism that essentially dilutes the expression of the transcription of the gene by a factor of two, could be some shared factor between the two new copies.


Arjun: And that would allow for essentially like a just in time down regulation of transcription right at the point of division.


Arjun: So I told my postdoc and my postdoc said, well, that's a crazy idea.


Arjun: This is Gautham.


Arjun: He's like, you know, that seems really crazy and sort of flies in the face of all kinds of different things.


Arjun: And so, you know, what is an alternative simple explanation?


Arjun: And I should say Gautham was the master of taking fun ideas and making them boring.


Arjun: I mean, he was like so good at finding these alternative hypotheses.


Arjun: And I asked him, I was like, I don't know you think of it.


Arjun: He said, well, I can't think of it right now, but you know, it takes time and hard work and you have to spend time on it.


Arjun: And I took that challenge to heart.


Arjun: Actually Gautham challenged me in many ways on that paper, but this was one where, you know, I sat down and I thought about it.


Arjun: And a few days later, I was like, you know, there is an alternative explanation for this.


Arjun: But you know, then I was like, okay, well, is there an experiment to eliminate that possibility?


Arjun: And you know, Olivia and I thought of what that experiment would be.


Arjun: And then she did the experiment.


Arjun: It came out that the interesting thing was actually the case, that there was probably some like just in time downregulation of transcription, which is I think one of the cool findings from that paper.


Arjun: But you know, I felt really happy about it that I had excluded a plausible alternative.


Arjun: And that is something that I was not well equipped to do before or had not thought about.


Arjun: And you know, it's boring thinking work.


Arjun: It's like work that you don't want to do is how do I disprove my hypothesis?


Martin: Right, yeah.


Martin: Popper would love you for that statement.


Martin: I reread that paper in preparation for our talk today.


Martin: And what I found amazing about that paper is that in the results section, which is very long, like every second paragraph starts with a question.


Martin: You know, when you read that, it seems like you're telling the story about how the project unfolded.


Martin: Is that really how it developed?


Martin: Or can you tell us just something more generally about the Night Science aspect of that project?


Arjun: The genesis of this project was very much a Night Science thing.


Arjun: Way back in Sanjay's lab when I did this RNA FISH stuff for the first time, I saw a lot of variability from gene to gene.


Arjun: But at the beginning, for the longest time, we'd always been looking at trans genes.


Arjun: And some of these trans genes just had high variability for whatever reason.


Arjun: And towards the end of my time in Sanjay's lab, we developed a robust method for RNA FISH for endogenous genes.


Arjun: And, you know, I looked a few and you'd see some variability.


Arjun: But as I started doing more of that, one of the things I realized is that, you know, we'd be counting per cell and that's what everybody was doing it.


Arjun: That was kind of the quantitative single cell biology du jour was, you know, count the molecules per cell and build models out of that.


Arjun: And, you know, we were doing that just like everyone else.


Arjun: But the thing that always bothered me is I said, you know, the thing is, you know, when I look in the microscope at a field of cells, if I could just take away the boundaries of the cells, so I wasn't counting the dots, but I was just looking at this field of cells, you wouldn't be able to tell where the boundaries of the cell are.


Arjun: And what that means is that the concentration of these RNAs should be kind of similar between the different cells.


Arjun: The cells are varying in size all over the place, but they keep the concentration pretty similar.


Arjun: And I just kind of tucked that little nugget into the back of my head, but it always would come back to me over time.


Arjun: And then finally, Olivia, so she was a physics graduate student, joined the lab.


Arjun: And as a quantitative person who's interested in these kinds of things, I was like, hey, you want to, shall we explore this some more?


Arjun: And she bit and we started working on it.


Arjun: And from there, it actually, there was a lot of night and day science in this story.


Arjun: I at least view this as very much a night science thing, that you make some weird observation, and often it sits in your mind for years before you do anything about it.


Arjun: And that was very much a case.


Arjun: But then, you know, it kind of faced the harsh light of day where we have to actually say, okay, well, what does that mean?


Arjun: And from there, you know, really had to think like, OK, well, you know, what do people know about this first off?


Arjun: And of course, you know, Paul Nurse had been working on this since the s and, you know, what can we contribute to it?


Arjun: And I think also one of the things and frankly, this is still always a struggle for me.


Arjun: This is something, again, that Gautham really made this point.


Arjun: He said, it's very hard to learn anything without making an invasive perturbation.


Arjun: Like, how are you really going to show something unless you really push the system in some way to reveal its inner workings?


Arjun: And actually, I would say that to some extent, the project did unfold in many ways as described in the paper, surprisingly enough, because that is most certainly not always the case or almost ever the case.


Martin: Yeah, it's rare.


Arjun: Somehow in this case, it did actually end up being that way.


Arjun: So that was actually, I think, the origin of what I think is one of the coolest experiments in the paper.


Arjun: So the basic idea behind the paper was that, you know, well, cells that are bigger have more RNA and cells that are smaller have less RNA.


Arjun: And the question is, does a cell actively sense in some way its size in order to produce the right amount of RNA?


Arjun: So how do you really show that this happens?


Arjun: And to me, the coolest experiment was, and this was a challenge that Gautham posed implicitly, is, you know, how do you actually show something?


Arjun: I said, okay, well, I'll show Gautham.


Arjun: Or rather, Olivia will show Gautham.


Arjun: So I was like, okay, let's do an experiment where we fuse together cells.


Arjun: So take a big cell and fuse it to a small cell and see whether the small cell, like if you add more essentially cell to it, will it produce more RNA to basically match the concentration?


Arjun: And something like that happens, and there are more subtleties to it.


Arjun: Well, you know, it's kind of night and day science.


Arjun: So here's a challenge, and ultimatum, can you really prove something, which is very much a day science thing, like you have this hypothesis, prove this hypothesis.


Arjun: Then it's kind of night science, like let me think, well, how could I actually do that?


Arjun: What's a cool way to do that?


Arjun: And then there's analysis of it and the examination of alternative hypotheses.


Arjun: And that's again, a day science thing.


Martin: It is amazing how diverse the things are that you do in that paper experiment.


Arjun: Right?


Arjun: It's like, you know, there's like these hetero-carry-on experiments and like all these.


Martin: It was a lot of tricks.


Arjun: Yeah, a lot of tricks.


Itai: A whole bag of tricks there.


Martin: Yeah, absolutely.


Arjun: Yeah, you know, it's actually, it's funny.


Arjun: So Gautham, he actually came up with a whole class of models that would be consistent with this data.


Arjun: And at some point I was like, okay, this is getting a little esoteric.


Arjun: So you kind of cut it off at some point.


Arjun: That's, you know, yeah, yeah, yeah.


Arjun: I mean, there is, you know, there is no end to the set of alternatives at some point.


Martin: Yeah, there's also no natural end for most projects, right?


Martin: I mean, there's always one more thing you can do to understand things better.


Martin: So I think one of the art of writing papers is to know when to start writing and to stop doing experiments.


Arjun: Yeah, I'm really bad at that, actually.


Arjun: I mean, I literally, we just have a paper right now, and I think I was, you know, my student and I were kind of lamenting that, like, important parts of that paper are lost because we just kind of put too much in it.


Arjun: And, you know, people's attention spans are like, you know, don't even read to the end of the tweet most of the time.


Arjun: So, you know, as much as I'd like to produce these beautiful scholarly works that stand the test of time, which I do think has value.


Arjun: But then, you know, we're in a marketplace of ideas.


Arjun: And I think I definitely suffer from having papers that are kind of too long.


Martin: Yeah, yeah.


Martin: I mean, one sentence that my postdoc advisor liked to say, and I'm sure you heard it too, Lawrence Hurst, in my case, one paper, one story.


Arjun: Yeah, there's a ton of truth to that, you know.


Itai: Arjun, I want to follow up on this idea of the marketplace of ideas that, you know, I want to ask you, how do you distinguish the not-so-great ideas from the great ideas?


Arjun: So it's certainly something that I've gotten better at over the years, is figuring out, okay, this idea's got legs or this one doesn't.


Arjun: I think I used to very much be like, oh, that's cool.


Arjun: It's tricky, it's clever.


Arjun: And it's sort of like, you know, this is my latest shiny toy I got.


Itai: Yeah, this is so cool.


Arjun: Yeah, exactly.


Arjun: And then the thing that I was not so good at was thinking, so what, you know, what's the next step?


Arjun: Where, how does this play out?


Arjun: Is it worth spending your time on really?


Arjun: And I don't know, I would say that what I've gotten a little better at is gaming out what a clever experiment or trick as it were, what it might tell us and how far we could take that.


Arjun: What's the next experiment that we would do?


Arjun: What are the other ways that you could do this?


Arjun: And really pushing people to think about other ways that we might get at it.


Arjun: Actually, and there is some systematic ways of thinking about this.


Arjun: Oftentimes, you know, this comes up a lot in the context of methods development, like, oh, this is a clever method, this is a cool method.


Arjun: And one of the things that I think is useful, you can actually systematize methods development, I think, by asking, well, what is the core innovation of what I'm doing?


Arjun: Is it just something that allows me to do something cheaper that some company will make cheaper in a year than probably not so...


Arjun: I mean, it might even be a clever trick, but it might not be so useful down the line.


Arjun: There are some systematic ways, I think, of evaluating that and really getting to the core of like, why is this clever?


Arjun: What is the key innovation here?


Arjun: What is something new about it?


Arjun: And then how might we leverage that in new contexts or here is a new place in which you could not do make an inference and now you can.


Arjun: So I've gotten a bit better at that.


Itai: I mean, the stakes are high, of course, because when it's the project of a postdoc and they're going to dedicate their next couple of years, at least to it, it's really, I think, a skill, right, to know of all the ideas that we've had so far, is this the one worth pursuing?


Itai: And different people have different style.


Arjun: Yeah, yeah.


Arjun: I mean, it's true.


Arjun: You bring up actually is sort of a tangential point is one about risk with people in the lab, because obviously, you know, they bear enormous risk in everything that they do.


Arjun: And it's something I'm very cognizant of.


Arjun: The one thing I found, though, is that, you know, boring projects take just as long and are just as risky as the fun ones.


Arjun: So you might as well just do the fun ones.


Arjun: I mean, unless unless the fun ones are, you know, sometimes it's fun because it's like, well, that would never work.


Arjun: In which case, you got to be a bit more careful about it.


Arjun: But, you know, I kind of feel like the thing to do now, perhaps this gets back to the question of how to evaluate things is just be patient.


Arjun: And wait for the project to come along that's like, this is fun and will work.


Arjun: And then just do that.


Arjun: So actually, a good friend of mine who I won’t name, so Uri Alon had this graph of, you know, how to choose projects.


Itai:  Yeah, right, in Mol. Cell, that same journal.


Arjun: Yeah, in Mol. Cell.


Arjun: And, you know, there are projects that are like really easy with small impact.


Arjun: There are projects that are easy with large impact and there are projects that are hard with large impact.


Arjun: And he's like, you know, you might want to assign this project to this person, this project to that person.


Arjun: And then a friend of mine said, he's like, duh, I mean, if I knew where every project was, I just aim for the top left every time.


Arjun: Like, of course, I just do that.


Martin: Boom, boom, boom.


Arjun: He's like, the problem is, you don't know.


Arjun: So you have to sample.


Itai: And that's the risk.


Arjun: I don't know.


Arjun: I kind of feel like, you know, there's very little point to doing things that are just not interesting, because in my experience, like sampling how hard it is, is pretty hard to gauge ahead of time.


Arjun: So just try and do things that are interesting, and sometimes it's easy, sometimes it's hard, sometimes it falls flat, and that happens, you know.


Arjun: The nice thing about academia, as opposed to like a startup or something, is that, you know, even if something doesn't quite go as planned, you can still usually publish a paper out of it.


Arjun: So we mitigate our risk that way.


Arjun: But I think that it's just hard to know what's going to work or not.


Arjun: Sometimes you just got to do it.


Martin: Yeah, yeah.


Martin: Actually, Arjun, I would like to go back a bit more to your Night Science techniques, right?


Martin: And initially you started by saying, well, maybe it's just a list of tricks, right?


Martin: And then the artist just to decide which trick would work in what situation.


Martin: But I think that's probably, well, maybe too modest, right?


Martin: Do you have any other conscious methods or habits that help you to come up with good ideas?


Arjun: Well, for me, I find that some number of good ideas, I'm a talker, as you might have guessed.


Arjun: I like to talk to people all the time.


Arjun: For me, I really get ideas and kind of pushed into new directions by talking to people.


Arjun: What I find is that I'm not a very good deliberate thinker, like sit down, think about something.


Arjun: I tend to work in bursts, so it's like I'll either have the idea quickly or I just won't ever have it.


Arjun: And part of the way that I get out of these local minima is to talk with someone and that I think just jiggles things up a bit.


Arjun: They say something new, fresh, different way, something or other, and then it triggers a thought in me or triggers a thought in them that leads to some quote unquote creative mew thought.


Arjun: There's actually another interesting trick that someone told me once, which is ask yourself what a creative person would do.


Arjun: It's surprisingly effective.


Itai: So who do you channel?


Arjun: I think just in the abstract, like, you know, if I was more clever, like what would I do?


Itai: What would Einstein do?


Arjun: Yeah, exactly.


Arjun: You know, I think actually, so this is another thing I personally think about creativity is that it's just a matter of energy.


Arjun: I think on some level, it's just how much time are you willing to devote to thinking about something in different ways or just, I don't know, keep pushing yourself in a way and always try and expose lazy thinking on your part and keep trying to think like, what's the core of what I'm getting at?


Arjun: What am I really learning?


Arjun: And I feel like that, if you keep pushing on that, that is another wellspring of creativity, which I think is again, just, you have more shots on goal, you have more chances to apply your bag of tricks to some set of problems, just the more you think about it.


Martin: Yeah.


Martin: No, I think that's a very valid point.


Martin: And like you translated that now, if you just imagine, okay, what would I think if I were more creative?


Martin: That's really just another way of saying, come on, just think hard about this, right?


Martin: Yeah, exactly.


Martin: So you're just taking the time to really think about it.


Martin: And I think that's actually a very important point.


Martin: I mean, just taking the time to think.


Martin: I have a question regarding your first part of the answer, which was that you like to talk to people and that helps you to be creative.


Martin: I'm just curious, do you typically talk to just a single person working on that project?


Martin: Or is that typically when you're with a whole group or like in what situations do you have these conversations?


Arjun: So I have a lot of, you know, walking conversations with people in lab and every so often that'll turn into some new idea.


Arjun: Sometimes it's in like a group meeting or meeting with a couple of people in a new collaborator or an old collaborator.


Arjun: So I don't know that there's any particular rhyme or reason to it.


Arjun: But I think actually though maybe I've had more good ideas in small groups of people than just purely one on one.


Itai: And you know, Arjun, also about this notion of setting aside time to think that we were just talking about, I think what's interesting about that is that, you know, these days we like to make lists, like to-do lists, and it's very easy to say, okay, I have to reply to this person, check.


Itai: But then to put on that to-do list, think hard about gene expression bursts.


Itai: You know, you think, oh, you know, why should I do that?


Itai: There's zero chance or only .% chance that anything fruitful will come out of it.


Itai: And so it's very likely to not get prioritized.


Arjun: You know, there's an alternative hypothesis here actually.


Arjun: Talking to people is not actually beneficial in any way about sitting in a room.


Arjun: It's just that if I talk to people, I forced myself to essentially be in a space where it's okay to think about an idea.


Arjun: Whereas, you know, if I sit in my room, just like, oh, well, what's new in my email?


Arjun: Turns out I have , forms.


Martin: Yeah, or you have to edit a paper or I mean, there's always something that you really should be doing, rather than just sitting there and staring out of the window thinking.


Arjun: Exactly.


Itai: And you know, also, when you look at your list of publications, it does mirror what you're saying that you talk to a lot of people and that these discussions are leading to a lot of fruitful collaborations.


Itai: I mean, you're extremely collaborative.


Arjun: Which is another way of saying I don't actually know anything.


Arjun: So that's why you have so many collaborators.


Itai: No, no, I think to me, it's really impressive, because I see that you're definitely leading so many projects yourself, but you're also a good sort of citizen and a good collaborator.


Itai: So I'm wondering, like, how do you do it?


Itai: How do you find yourself in a collaboration and what's that dynamic like?


Itai: What are the night science aspects of it?


Arjun: Yeah, you know, it's, I mean, collaborations take so many different forms.


Arjun: Some of them are very much day science kind of things where it's like, we need to prove this point.


Arjun: You know how to do this experiment.


Arjun: Can you please help us do it?


Arjun: You know, and, you know, that's part of doing things.


Arjun: Night science collaborations are, I find them much rarer where you have someone with whom you can just talk about random things.


Arjun: And, you know, it's interesting because I actually kind of wonder if those collaborations are less likely to lead to co-authored publications.


Arjun: It's collaboration in a broader sense of just like, I love to have ideas with this person.


Arjun: And just talking with them just gets me thinking about all kinds of fun things.


Arjun: And maybe I'll turn it into a project and maybe that person is like, quote unquote, on the paper or not, but, you know, I've also had, you know, actual legit on paper collaborations with people with whom we have had very, you know, like night science discussions about, you know, what could this mean?


Arjun: What does it mean?


Arjun: I wonder if to some extent night science is a bit solitary in that regard.


Arjun: I don't know.


Itai: Or solitary in that it's intra lab.


Itai: In other words, the day collaborations are between labs, but to really do night science, it's maybe someone that you have a more intense relationship and you talk to them every day, you go to lunch with them.


Martin: Yeah, I think it's a very interesting question.


Martin: But my suspicion would be that in theory, doing night science with another scientist from another lab that has a different background and has a different list of tricks could be extremely productive.


Martin: But maybe we just don't do that because, you know, it's not expected, right?


Martin: We don't go to them and waste their time by, you know, just telling them what's going through our mind if it's not fully formed.


Martin: Maybe it's just a barrier like that.


Martin: I don't know.


Arjun: You know, that's a really interesting point, because I think about people with whom I've, you know, worked with where it has been about ideas.


Arjun: And yeah, you have to be able to communicate in a way that is very informal to the point where it's like, okay, just suspend your rational brain for a minute.


Arjun: Like just imagine you could do X, Y, Z.


Arjun: Like what would happen?


Arjun: What could it tell you?


Arjun: Like, oh, that's cool.


Arjun: A free association that, you know, frankly, you just don't have time to build that kind of relationship with most people.


Arjun: Yeah.


Martin: I think that's true.


Arjun: I think you're onto something there.


Arjun: I think that really is true.


Itai: Build some trust.


Itai: Yeah.


Martin: And, you know, that was really interesting what you just said, you know, that you have to feel free to some extent to even suspend rationality.


Martin: I mean, to just freely think about things.


Martin: I mean, one thing that's noticeable is that, you know, when one sees what you talk about in public, one topic that comes up quite a few times is free will of cells.


Arjun: Oh yeah.


Martin: Do cancer cells have free will?


Martin: Which is a beautiful way of looking at them.


Martin: Is that just a way to explain to non-scientists what the kind of questions are that you think about?


Martin: Or is this kind of anthropomorphism actually something, some technique that you use in your night science?


Arjun: There's a lot to that actually.


Arjun: So now I would say, now that I think entered a phase in my career, I sort of have night and day science on my mind all the time.


Arjun: It's been very interesting because I tend to I think, be much more philosophical and definition driven in how I think about science.


Arjun: And as a consequence, I don't know, it's really like I view definitions as progress and you have to think hard about them, but then how they fit together is like a piece of art, right?


Arjun: It's like a puzzle and you have to sort of master the technical parts of this puzzle, but ultimately they come together into something beautiful.


Arjun: Like it's these definitions need to somehow lock together in an aesthetically pleasing way.


Arjun: And I think that is sort of where I'm at now is really trying to unite these sort of night and day things.


Arjun: Like come up with things that are hard and rigorous, but that have an aesthetic beauty to them.


Arjun: It's like, we're all in a way artists doing basic science, right?


Arjun: We're artists trying to carve a sculpture out of this raw block of empiricism that the world is.


Arjun: Like we're trying to make sense of it in some way and to give it form and shape.


Martin: And to find the statue within.


Arjun: Exactly, yeah.


Arjun: And how you sort of set up those problems, that's I think our form of art.


Arjun: And it's taken a long time for me to really fully appreciate that aspect of it.


Arjun: I think in biology, it's, you know, everyone says of biology so stubbornly resistant to definitions.


Arjun: And again, I think that's just a matter of us not thinking hard enough.


Arjun: I think I'm not saying that I have the answer, but I see a path at least for myself and how I think about, how I think about that science, how to use definitions and how to see the art in the rigor.


Arjun: If that makes sense.


Itai: No, it makes perfect sense.


Itai: That was beautiful, Arjun.


Itai: And I think the sense of art and the sense of rigor are really nice kind of analogues to thinking about night science and day science.


Arjun: Yeah, it's just, kind of amusing about is like, yeah, I think we also we live in this era of insane technological progress.


Arjun: Even in our science, right?


Arjun: Everything is so technologically driven.


Arjun: I think there's this kind of Silicon Valley mentality that technology will solve everything.


Arjun: And I think there's this idea in the parallel in science is this idea that data will tell all, don't bother thinking.


Arjun: And I think that the analogy goes the following.


Arjun: So if let's say we live in this world in which technology has solved every problem, at the end of the day, we're still humans, right?


Arjun: We're still humans living in this world.


Arjun: In the end, we would just be artists.


Arjun: Like if robots did all of our work for us, did all the things that we don't want to do, all that's left is art, that's what humanity would be.


Arjun: And we get some small sliver of that in our real lives.


Arjun: And I think it's worth remembering that.


Arjun: And I think in terms of how we do our science, it's the same that we need to think like, what are the questions that got us excited about?


Arjun: What's the art that we're going to make?


Arjun: And, you know, yeah, we have to do all, you know, take all the data, do all of these things to do that.


Arjun: But it's important to take at least some time to think about the art of it, because at the end of the day, that's all that's left.


Martin: I think that's a wonderful way to wrap up our discussion.


Itai: Night scientist as artist is definitely a notion I can subscribe to.


Itai: Arjun, thank you so much for taking the time to talk to us.


Itai: I speak only for myself, saying that it was just really inspiring.


Martin: Well, of course, I cannot but agree.


Martin: It was really amazing.


Martin: I mean, there's a couple of things that I have to try out in my own Night Science now, I think.


Arjun: Well, the Marketplace of Ideas continues.