Phase Space Invaders (ψ)

Episode 22 - Lucie Delemotte: Enhanced sampling methods, alternative publishing models, and becoming a parent in academia

• Miłosz Wieczór • Season 3 • Episode 22

Share episode

Send us a text

In episode 22, we start by talking about the 2022 review of enhanced sampling methods that Lucie co-authored, one that provides long-needed organizing principles and unifying vocabulary for applications in this field. We also outline some challenges and community needs still waiting for the bold souls out there looking for ambitious projects. Lucie then moves on to share her experience with alternative publishing models, both as a member of Biophysics Colab and a former editor of eLife, and we dwell a bit on the challenges and possible solutions to what some would call a crisis in publishing, where evaluation metrics and publishing companies created unsustainable or even counterproductive incentive structures in sharing scientific results. We end on personal notes: already last year when I was visiting her lab in Stockholm, Lucie agreed to be on the very first list of interviewees, but it took me so long to figure it out that eventually Lucie gave birth to her twins and went on a maternal leave, so now that she's back more than a year later, I also asked her to share a few reflections on becoming a parent in academia.

Milosz:

Hello, this is the Phase Space Invaders podcast you're listening to. We're at episode number 22, and today I'm talking to Lucie Delemotte, professor of biophysics at KTH Royal Institute of Technology and SciLife Lab in Stockholm. Lucie's lab has been focused on the conformational dynamics of ion channels and transporter and receptor proteins, running many successful projects in the subfields of voltage sensing, ligand dependent activation of GPCRs, or modulation of activity by lipids. All these applications are tied together by the need for enhanced sampling approaches, or ways in which we can accelerate the statistical exploration of molecular geometries that are relevant to physiological function. And so we spent a good while talking about the 2022 review of enhanced sampling methods that Lucie co authored. One that provides long needed organizing principles and unifying vocabulary for applications in this field. We also outlined some challenges the community needs that are still waiting for the bold souls out there looking for ambitious projects. Lucie then moved on to share her experience with alternative publishing models,both as a member of Biophysics Colab and a former editor of eLife. And we dwell a bit on the challenges and possible solutions to what some would call the crisis in publishing where evaluation metrics and publishing companies created unsustainable or even counterproductive incentive structures in sharing scientific results. We end on very personal notes. When I was visiting Lucie's lab last year, I was pitching the idea of this very podcast to Antoni Marciniak, my former mentee and now Lucie's PhD student. And Lucie agreed to be on the very first list of interviewees, but it took me so long to figure it out that eventually Lucie gave birth to her twins and went on a maternal leave. So now that she's back, I'm After more than a year, I also asked her to share a few reflections on becoming a parent in academia. Hope you have as much fun listening as I had talking to Lucie. Let's go! So Lucie Delemotte, welcome to the podcast.

Lucie Delemotte:

Thank you. It's a pleasure.

Milosz:

Lucie, aside from your applicational focus on the regulation of membrane transport, so ion channels and transporters, you've spent an impressive amount of time thinking deeply about enhanced sampling methods. And of course, this subfield has such a rich history and, you know, so many genius thinkers devoted decades to coming up with ever newer approaches to this idea of more efficiently extracting information from molecular systems. their tens of thousands of degrees of freedom. But all that work, do you have your favorite one? And which one is that? If you, if you care to tell us.

Lucie Delemotte:

Uh, yeah, I think in fact, I would say I don't have a favorite one, because the whole sense of the review we recently published was really reviewing what exists and making sure we know what method to use in what context. And it was quite enlightening reviewing all of this to also realize that, you know, some methods are popular for the wrong reasons and some, and the other way around, I think. So it's a really kind of complex question where the availability of code, the community that uses it. The visibility of the method, even picking a good name, might impact how much a method ends up being used or not used. So I don't think we always do this for the right reasons. That was one of the main aims, actually, of that review, was to make sure that we educate people. First, I wanted to educate myself. That's the real reason for writing the review, but I thought it would be useful to just spread this to the wider world and make sure that people are at least aware of what exists and decide what to use for good reason. So depending on, you know, how much information you know about your system, how you have structures, are there of good quality? Is the process that you're looking at mostly on a single path, or is it a redistributed process? You would make different decisions.

Milosz:

Right. And for the people who are working on the next method, do you still think we need more? I mean, there

Lucie Delemotte:

Yes.

Milosz:

cases where we are still lacking

Lucie Delemotte:

Yeah, can I just, Uh I'll just say how this idea, how this review was born, because I think this will answer your question directly so now this goes back quite a long time. At the end of my PhD in about 2010 11, I reached the point where I was working on these ion channels you mentioned, and I reached the point where I wanted to simulate the whole activation process of a given ion channel, and, it just wasn't possible with, regular molecular dynamics simulations. And that's actually still true today, but this being, uh, more than almost 15 years ago now, this was really not possible at the time. So I started looking into enhanced sampling. And the only thing I, knew about at the time for various reasons, Was metadynamics. And that's kind of why I was joking about the method having a good name, because I think it was kind of genius to call it that way, So I knew about metadynamics and I figured we can just, uh, use that. And, uh, you know, at that time I probably wasn't even aware of the problem that there was the method that you choose, but also the. Whether if it is a collective variable based method, then we should really be careful about the choice of the collective variable. So I launched into this without knowing very much about it. And of course, it was disappointed because it was very naive in the first place. It didn't work and, you know, I spent really years trying to figure out how to use this in the best way. And. me being the person I am and just probably talking too much. I was going around conferences and like workshops and being all disappointed about it and asking people for feedback. And, I remember I really wanted to acknowledge here, Jérôme Hénin, who's My coauthor on the review, because, uh, I remember very distinctly early conferences talking to him and asking him, could I use your method, which was, which is adaptive bias ABF adaptive,

Milosz:

By

Lucie Delemotte:

I think fourth method. And he was telling me, yeah, sure. You should use it. And then, you know, I was talking to other people and, you know, they were telling me I don't know about that method, but you know, my method is really better, you should use it. And, uh,

Milosz:

see where

Lucie Delemotte:

So it was really like this and you know, I was still relatively young and, also full of doubt and I was like, ah, they're probably right. This should work. And then, you know, things were not really working for me. And then fast forward, I gained a little bit of assurance, you know, I wrote papers, I got a job. Or several different jobs, including a tenure track one. So I was starting to think, maybe I'm not that dumb after all, because I am apparently able to convince people that what I'm using, what I'm doing is useful. But I really don't, still don't know what enhanced sampling method I should use, or what type of collective variable I should use. And so in 2018, we were at this, workshop on protein dynamics in Lesouches, in France. And Jerome was there, and, again, like, poor guy, I was again bringing up this question of, like, what method should we use, and what exists, and stuff. And I discovered that this guy had a whole table in his computer where he had listed a, you know, a whole very large number of references. referencing different techniques. And it was not an Excel sheet because he's really a Linux type of guy, but I forget what it's called. Open the open version of Excel at that point. And it also listed, you know, some characteristics of the method and in which way the papers were different. I was like, my God, this is genius. We need to turn this into something that is useful for other people. And then we decided, I decided that it would be a good idea to bring this to life. But really, I felt still, I didn't know that much about it. It was really Jerome who had all the knowledge, but I thought that I could help write it. And he agreed, actually, that in explaining things to me, Then we could write it in a way that, you know, I'm the target audience. And so other people like me are the target audience. So I felt like, you know, I was a bit of a midwife in that process and like bringing to life Jerome's, vision.

Milosz:

The Socratic method.

Lucie Delemotte:

Yeah. And. In fact, this is 2018. And, in a way for us then the Covid pandemic was, you know, a good, time because we could spend a lot of time working on this because basically many other things were shut down. So it was a good time to write this. And other people joined us, long time because we started talking about this at conferences. And so since it was a big effort, it was good that our other co authors, uh, joined us, but, uh, we spent a lot of time during COVID meeting sort of weekly about this. And, uh, it was a long process. So it took, I think from the beginning to end. When was this published? 2023, maybe? So five years to put together. It was a very long process because busy people and, um, and it was a hard job, especially the hardest part, I think, was that we decided to review the methods. And Write the equations in a consistent way because I thought it was hard also is that different papers were using really different Ways of writing them so that took a lot of work to Do in a consistent way. So sorry, this is a very long Explanation, but I forget what your original question was. But oh, yeah, I remember now. Do we need new methods? So, I think, I think it was funny because in reviewing this, we realized there were really only a few. either physical or statistical principles that were used, and they were just recycled over and over again and rediscovered and called something new. but we thought by the end of reviewing this that the area where there was really promise was where these principles were leveraged and combined in an advantageous way. So, for example, you know, it's a popular way to increase temperature, but it doesn't solve all of your problems. And it's a popular way, I don't know, to sample along a path. But if you can combine high temperatures in a smart way and the path, then you might be able to really enhance the sampling. And then, of course after this there was the sort of sort of, we can call it, I guess, AI revolution, machine learning. So, we didn't really launch much into describing this, but, there's also. Promise to accelerate many, many aspects of these enhanced sampling methods piecewise with data driven approaches applied to these different parts. So, that we didn't review very much because we came, I guess, right when this was becoming very popular so we don't need new

Milosz:

abbreviations.

Lucie Delemotte:

Was that? Yeah. Yeah, that goes back to finding good names, I guess But ultimately, I'm, it's good because it took such a big effort, but this paper is, has become really popular. I mean, compared to other things I've done, it's like becoming the most cited review. So I think it was really responding to a need that people had

Milosz:

hmm.

Lucie Delemotte:

in the community. I'm very glad we were part of helping with that.

Milosz:

Out of curiosity, have you had a chance to test out at least the majority of the methods described?

Lucie Delemotte:

No,

Milosz:

Or all?

Lucie Delemotte:

no, no, no, no though. I mean, it's been a while. There's a whole like all these, adiabatic, uh, methods that I remember testing in my PhD, but like for a week or something and just being like, I really don't understand how this works, so maybe I forgotten that I've tested. Many things, but now basically we use a lot AWH and that's because, it's, natively in GROMACS. So we find that it's really convenient to have a method that is faster just because it's really in GROMACS directly. And, still when we don't know much about the system, I find that, replica exchange methods can be very powerful. So it really depends, I think, on the, on the type of problem, but, uh, a lot of the work now is AWH. And again, I'm not saying that it's because it's a better method than another one. It's really for convenience reasons and speed.

Milosz:

Right. I think the adoption in software is really a big driving force, right? Jerome himself is developing Colvars, if I remember

Lucie Delemotte:

Yes.

Milosz:

And we've had Max Ponomi here on the podcast who had his role in developing Plumed.

Lucie Delemotte:

Yeah,

Milosz:

And of course, each code incorporates its own approaches. Gromacs has, as you say, AWH.

Lucie Delemotte:

yeah, exactly. And, Colvars is, originally, it's born in the NAMD community, but, they also actually have an interface to GROMACS. we haven't tried to, to use it much, so I don't know how it compares in terms of speed. hmm. Mm

Milosz:

yet, but at some point I will probably do that. Um, Yeah. Do you think there's a sort of trade off also in those methods between being perhaps elegant and efficient? Because everyone, of course, tests their methods on the simplest cases, right? There's always this problem that, okay, we have the alanine dipeptide tested 1 billion times with every possible method, and it always yields the perfect, converged, energy surface and some methods have this appeal of being like exact in some limit and the swarms of trajectories for example but yeah sometimes they are just working great on a small problem and don't really scale well to larger real life problems right

Lucie Delemotte:

Yeah. So I, first of all, I think the community has made a lot of progress. So we were always joking, few years ago that, Oh yeah, it works on alanine dipeptide. Great. I think it's become different. So I think now it's become, frowned upon if you publish a paper that just has alanine dipeptide as a reference case, but it does speak about also how do we validate these methods? Because often we're trying to enhance the sampling because we cannot actually reach at least the free energy, and even worse, the kinetic properties, without, enhanced sampling. So we, often we don't have the ground truth to compare to. And I think that's why alanine dipeptide is still popular is because we have a exhaustive sampling of this in the, in the introduction. Our applications are really quite complex. I mean, ultimately, our group is really interested in pretty complex, things, not really developing methods, but rather, investigating objects like transporters and regulatory modules of ion channels. And then for us, we haven't been after the sub-kJ per mole type of accuracy because, we know we can't reach that type of thing. So if we have a semi quantitative whatever that means if you know, we have at least the right direction of the free energy, let's say a state is more stable when the ligand is bounded and it's as expected. I think we're already quite happy with that sort of result. So the focus recently has been a lot more on finding Typically collective variables that allow us to describe, the process and the free energy method has mattered a lot less, but in my discussions with Jerome and again, I'm not an expert. So really you should invite him. I think if he hasn't been on the podcast, I think he's a,

Milosz:

a good

Lucie Delemotte:

he's a, he's a really, um, great colleague for me. I think I've learned so much in talking with him. but, uh, in any case, he has. said that, when you have a better control over the process that you're describing, then the free energy estimation method that you use, becomes important because you know, it will be more or less accurate. And then you might want something that is so called more elegant, but our focus has really been on sort of more the efficiency side in the sense like. We're still looking at quite approximate things. And so we haven't been so anal about finding the estimation method that gives the most accurate results. Hopefully we get there one day.

Milosz:

Right. I was talking to the Plumed guys one day and, they pointed I don't remember if that was Max Bonomi or Giovanni

Lucie Delemotte:

Mm hmm. Mm

Milosz:

that we don't have good benchmarks for free energy methods. And, I think I suggested at that time that, you know, there are some systems, you should be familiar with them from the transporter side, which have inherent symmetry, right? So there are processes that should

Lucie Delemotte:

hmm.

Milosz:

Delta G of 0 by construction, some sort of flips in the membrane that preserve some sort of symmetry. So maybe we should think in those directions about,

Lucie Delemotte:

but what do you mean? Sorry, now I'm interviewing you

Milosz:

uh, no, if you have something that for example, opens to the extras versus intracellular side and is a homodimer, right the whole flip, the conformational change could have a delta G of zero,

Lucie Delemotte:

but, uh, at least with the transporters that we work on, uh, usually it's, uh, it's a zero sum game in terms of the confirmational cycle, but there is a more stable state typically open to the outward, if it's an importer, right? It's open to the outside and then when the substrate binds, it sort of catalyzes the conversion to. The inward state.

Milosz:

there are very few that have this property that are symmetric. I remember working on one of them back in the day,

Lucie Delemotte:

Okay, I see what you mean.

Milosz:

still find them. So I'm thinking, yeah, if we can people to look for more and more complex cases to validate that, I think the next stage that would be super exciting would be actually to take all those methods to kind of optimize them and see. which one is efficient in specific cases,

Lucie Delemotte:

but I think in

Milosz:

if,

Lucie Delemotte:

past, maybe three to five years, I've been to a large number of workshops where this has been brought up. We need a benchmark set and, CCAM meetings, but also other types of meetings sort of adjacent to that. And I think in many of these workshops, where people like Gerhard Hummer. and we kept coming back to what would be a good benchmarks

Milosz:

yeah.

Lucie Delemotte:

many times. And somehow this never materialized into anything just because it wasn't an

Milosz:

because

Lucie Delemotte:

obvious answer. I, we didn't agree. And Gerhard himself, I think, has tried in the papers where they're doing this. To come up with a subset of, I think, so there's often an ion pair solvated in water, and then two helices in the membrane and, you know, typically the T4 lysozyme, I don't know if I'm saying something stupid, that's maybe not in their paper, it's, it's in different papers.

Milosz:

remember this

Lucie Delemotte:

you know, so there's a few test cases like this, and then other labs use really the type of problems that they're interested in. But somehow it's never been really absorbed by the community as, uh,

Milosz:

Yeah,

Lucie Delemotte:

the benchmark said. And I think it goes back to this ground truth I was mentioning earlier. We just don't have it for the interesting cases.

Milosz:

right, symmetry would be a good, I mean, I'm just throwing this idea out there. I'm not probably going to have time to work on this, but if someone does,

Lucie Delemotte:

You need to write a grant, Milos.

Milosz:

Okay, but maybe let's jump back into applications a bit more. So as we mentioned, you, well, you approach the world of ion channels and from the less mainstream side, which is more on the side of regulation and kind of biologically relevant conformational changes. Do you think there's something that, you know, miss from the standard, oh, iron goes through the channel picture that we get sold the textbooks? Like, what is the broader biological picture that you're after?

Lucie Delemotte:

Um, I think, okay, so I'll tell you the truth.

Milosz:

That's what we're looking for, ultimately.

Lucie Delemotte:

Um, and, uh, this is a bit humbling also. So, um, I always wrote, in typically grant applications. once we get the whole conformational ensemble for, you know, typically the voltage sensor domain of ion channels, then we can design modulators that will modulate this. And then, a couple of years ago, we, I think we had this breakthrough with, my student. then student, now researcher, Darko Mitrovic, where we found ways to basically, simulate those conformational ensembles. And then I swear I had bit of an existential crisis because I was like, what are we going to do now? I've been trying to do this for, you know, eight years or something. And I really had the month of like, what are we doing now? And I knew that the step should be designed because that's what I had been writing. And it seemed like a dream, but I really didn't have any idea on how to go about it. So even though somehow people think I'm a ion channel slash transporter person. I don't know if I have that much to say about the objects. It feels like our angle has often been to enable new things. Not really with designing methodologies, because I don't think we're really belong to that world either. But, you know, making the methods that smart people design, accessible, like workable and accessible to discover something about nature that will then become engineering eventually. But it's really the experimental collaborators then that come up with the interesting, uh, biological insights. So yeah, I don't know how I would really summarize what is cool to know about, about ion channels and what people miss about this class of protein. I think it's a bit true that they're just a hole with ions.

Milosz:

Okay.

Lucie Delemotte:

And then there's a lot of really interesting things when it comes to the details, but in terms of the broad workings of those,

Milosz:

Yeah, maybe the details are also interesting,

Lucie Delemotte:

the details are super interesting. So, you know, there's just little stories here and there that are really fascinating, but I don't think it changes the way the textbooks work.

Milosz:

I remember you worked on this hERG protein, for example, right? Which is so famous for having so many medicinal applications or implications for people who take different sorts of drugs, right?

Lucie Delemotte:

Yeah, that's because it's a cardiac channel, so it's pretty bad if you, hit it with a blocker. basically people will, die from it, so you need to be really careful with that one. Uh, but that's actually an interesting story because, the person who revealed the role of hERG in that context, Gail Robertson, she says that somehow the regulatory instances took a hold of this idea that it was really important to avoid hitting hERG and it became part of the panels that all the drug developers need to screen against. And she says now, I think that, this blew out of proportion. So there are many other targets that you would like to avoid and somehow hERG became really kind of important. I mean, not for bad reasons, but it's just that there's many other things that you should also be looking at.

Milosz:

see.

Lucie Delemotte:

So she was herself a little bit surprised that this happened, I think.

Milosz:

Well, so it's very bottom line is that we should pay less attention to the ion channels in the end.

Lucie Delemotte:

No, that cannot be true. But ion channels, uh, it's Chris Miller who was saying this. Ion channels are a little bit boring. Transporters are more interesting in terms of their dynamics. Like, they actually move more. That's why I got into the regulation of ion channels, because I'm more interested in the parts that move around the channel that made it, make it open or close, or even, you know, the lipid environment or something. The hole itself is just a hole.

Milosz:

No, absolutely. I think, I think a good rule of thumb for choosing projects in simulations that at least something should move

Lucie Delemotte:

Yeah.

Milosz:

simulation. And not in a sense of translation or rotation of the whole molecule.

Lucie Delemotte:

Yeah. Yeah. That's right.

Milosz:

should be happening.

Lucie Delemotte:

in my postdoc, in Ursula's group, I did, a little bit of QMMM and I watched a proton jump from one, from one residue to the next. And I was like, that's cool to see once in your life, but, this is not enough movement for me.

Milosz:

I agree with that. Yeah, I've seen that too is exciting for the first 10 minutes

Lucie Delemotte:

Exactly.

Milosz:

Okay, so moving on the next topic that we had some discussions on. Exactly we converse around the question of publishing models, right? Because there is now so much talk the legacy publishing models perhaps being outdated, perhaps being hijacked by corporations, perhaps not serving the community. I mean, people have different angles. So what's, your angle on that? You served on eLife's board,

Lucie Delemotte:

Yeah.

Milosz:

a while.

Lucie Delemotte:

Yeah, I don't even know really how to start, the discussion here, but, basically I've for a long time been very. Disappointed, I think you might say with the way we, we publish science. I've just seen really bad behavior from various, colleagues when it comes to publishing. And, you know, most of the Fights that I've had with, even collaborators ends up being when we decide to publish the papers and how do we write them? Who's an author? What place they have? Which journal do we submit to? Which reviewers we pick? And we waste so much time and effort and even emotional energy on the topic. So I think this is, as sort of everything in life, when it touches you personally, you start becoming more invested in the question somehow. but I, so I've been very interested in,, alternative models basically, and I was, I served on the eLife editorial board for a while because I thought this was a really modern venue. I mean, I was invited and I had no doubt in saying yes. to that. then I, I recently quit actually because I was becoming really too busy and I needed to find a way to reduce the amount of time I was spending. But in the meantime, I got involved with science collab specifically, which is a preprint review, and then curate, type of model. So it's similar to eLife, but it has. more agency because it's a smaller community of people and it's really focused on biophysics. And it's been really inspiring actually to be part of this and try to see how we can come up with a really, um, more equitable, inclusive and transparent model that, could help to get rid of the bad behavior that's related to publishing in journals. So this is not very concrete, but if you want me to say more concrete things about it, I will.

Milosz:

Well, absolutely always welcome. I think, well, one thing is that this is a sort of subculture within academia Right. now. Right.

Lucie Delemotte:

Okay.

Milosz:

cause I love. A lot of, I think, social change happens through subcultures promoting certain behaviors and then this thing's catching up, like the rest of the population catching up with them, but in academia it's hard because you're also being constantly evaluated, right, so your ranking is

Lucie Delemotte:

Exactly.

Milosz:

of tied to where you are in the mainstream culture, not in the subculture, but

Lucie Delemotte:

that's absolutely right.

Milosz:

it,, to popularize subculture so that it has some real life impact on the mainstream, Publishing culture,

Lucie Delemotte:

yeah,

Milosz:

but yeah, please go ahead with the specifics. I mean, I think

Lucie Delemotte:

No, but you're, you're, you're absolutely right about this. And I mean, I feel very torn myself about this because I feel like there are many, many issues with the traditional publishing industry. And I think a lot of people are aware, but basically it's not inclusive. It's not transparent it's slow because, all the papers get rejected. It's unfair in the sense that reviewers don't get paid and authors have to wait. And the people who are making the money are not the scientists themselves. There's many, many, many issues not to mention the thing that I find the, the saddest is, you know, the people who are the most, popular if you go to conferences are the editors at the Nature branch journals instead of being the people, you know that you would talk about your science instead A lot of people are just like chasing the editor because they do know that there's a human factor in it so if you Know people you have a higher chance of your paper getting accepted in it So this is all stuff that really makes me cringe, and I think make a lot of people cringe. And then is the other side of what you just described. It's really complicated because this is how we get evaluated. And now to me, it matters much less because I have a tenured job at a good university. in a country that is also kind of friendly. And if I don't get papers in these journals, the consequence for me is probably I don't get the next grants. And then I do less science, but I still have a livelihood and a job. But what about the people who are coming through my lab? How do I work with that? Because they need to move on to their next jobs too. And so the people who are set on staying in academia, for good reasons, they're insisting on being part of the mainstream, like you're saying, and it's hard for me to go against that, because I'm trying to be considerate of them also. And then you're a bit of a hypocrite, because you're, you know, taking part of a system that you're also fighting against. But it's very difficult to do otherwise. So I'm excited that there's awareness about this, that there's initiatives that, you know, we're trying to reform the way we assess research. There's big coalitions right now that are being put together. And there are indeed, Actors, different grant funding agencies, different universities that are trying to think about, this research assessment differently but it's a difficult problem. There's no, really good way to evaluate, especially the big entities, so you can evaluate a single researcher based on their research, and you don't really need even to look at the types of journals they publish in. It's harder when you're selecting between, I don't know, 50 candidates for a job, then it's really impossible when you're evaluating a department or an entire, institute. There's something you need measurable quantities, and then it's hard to not go to impact factors or something that is really, you know, data points. So people are working on this and it's good. I don't know how hopeful I am that we can really, Make changes in terms of research evaluation, but I'm still glad to be part of, like you say, the subculture of, I mean, maybe that's what a professor should be doing right once you have tenure is to dare to go into these directions. But the problem is again, the interest of the trainees that is directly linked to your own. So I think that's why change is slow. And sometimes it's a little bit discouraging but somehow it still matters that we do it just because, the current system is just so terrible. And in fact, it ends up hurting the actual trainees indirectly. So it's, uh,

Milosz:

And science also in some way

Lucie Delemotte:

Yeah,

Milosz:

a

Lucie Delemotte:

no. And humanity ultimately, right? Because we're, you know, it's slower to make progress in terms of questions that impact us directly. Now I'm thinking about, climate change and things, but in our field, even bringing new ways of modulating channels or transporters could be path to treating orphan diseases. So

Milosz:

Right.

Lucie Delemotte:

things should go faster if possible.

Milosz:

Now, I think one optimistic note on which I could think about it is that you're right to say, it's a very kind of hierarchical problem right on many, many levels of organization. And probably in the end, it's more up to us scientists to propose solutions that say governments will be happy to accept or adopt they're robust, like, it's not that some legislators will come and say, Oh, you know, we want to measure you by your Hirsch index. It's

Lucie Delemotte:

I think you're right.

Milosz:

come up with something that works and that everyone accepts. It's what people will be happy to evaluate us on. So maybe we're really missing some ideas there. And of course, some consensus.

Lucie Delemotte:

Yeah, but I think at least as far as I know, I haven't heard really about solutions for the larger entities. I think I'm very fond of these narrative CVs. And, I'm actually part of the Swedish Young Academy, and, you know, every year we elect new members. And, uh, this year I was very happy to be part of designing how we will, do the selection. And indeed, now, this year, for the first time, we're asking for people to, list their research outputs. And it doesn't have to be papers, so it can be computer code, it can be, artistic productions, it can be, whatever you think about. including, you know, outreach, maybe a podcast would be something that you could, but, uh, anyways, any sort of output like this, and then explain in which way this output is important and what it contributed to society or science even. And then how do you link these five outputs together to create a coherent profile? I think this is brilliant, and I, I think it's really nice to read those things as opposed to going into a list of papers that, you know, says

Milosz:

Oh

Lucie Delemotte:

56, this is the title, 57 is whatever title. I think it's, it gives you a much more precise idea of a person, but, I don't know how you do this on a bigger scale. Somehow, you have to condense All of this, but imagine a department with, I don't know, 20 professors and maybe 50 people total, how do you aggregate this now into you write a little text, but it's very difficult to convey. in which way this is, you know, better or worse than the next door department. So I don't know how to do this. I think it's complicated,

Milosz:

Yeah, it's not right.

Lucie Delemotte:

I agree with you. I don't think there's a higher force that says this is how you want to do it. It's just that we haven't really come up with a better solution as far as I know.

Milosz:

Yeah, maybe there is a multiple tier solution,

Lucie Delemotte:

Yeah, probably.

Milosz:

depending on which level of evaluation we're talking about, maybe it will require different tools, different numbers.

Lucie Delemotte:

That's right.

Milosz:

will be some numbers involved eventually, but

Lucie Delemotte:

Yeah, yeah, exactly. No, you're right. I think, uh, I think that's a bias of mine and of humanity in general. Maybe we try to look for a global solution, but in fact, here we need to be nuanced and decide that we can evaluate different things on different bases. But in any case, I think, participating in biophysics co lab and having a way for the community to evaluate preprints. in an open way and public and then provide A coherent set of, suggestions to improve the paper, to validate better the data, to support the conclusions and offering this to the authors and giving them the opportunity to respond in the open. I think this is, in any case, how we should do it, even at. Traditional journals because just that part of it, I think, would, go a long way to improve, how we're doing things, but I think it goes against some of the interests of the publishing industry. But,

Milosz:

Right,

Lucie Delemotte:

um, like, for example, at Colab, the curators who are basically editors, but we call them curators because it's not exactly a traditional journal and the reviewers get paid a nominal fee for their review. This sort of thing.

Milosz:

Right.

Lucie Delemotte:

how much it costs to publish, you know, a paper in Nature Communications, I feel like you could pay, I don't know, a hundred dollars to a reviewer, you

Milosz:

Oh,

Lucie Delemotte:

know?

Milosz:

That would be nice. That would be nice.

Lucie Delemotte:

That'd be nice. That would

Milosz:

will

Lucie Delemotte:

incentivize also doing it and doing a good job of it. I mean, how long does it take to review a paper? It takes a day or something, right? Like it's,

Milosz:

Yes.

Lucie Delemotte:

a substantial piece of work if you do it well

Milosz:

absolutely. It will also be interesting to see in the comment section under articles like we have on YouTube or other platforms,

Lucie Delemotte:

And, you know,

Milosz:

What do people have to say or question or even undergrads coming and asking, Oh, you know, what about this thing? I don't understand. Can someone explain, right?

Lucie Delemotte:

I think so. And speaking of the analogy to movie reviews and so on, I thought it was really smart, actually, what eLife decided to do, but that has not been really accepted,, broadly. And again, it's linked to all the issues with having impact factor and so on. But this idea of we don't need to just accept or reject the paper, I think, associating with the paper some small editor blurb, describing the quality of the paper and its impact with this kind of controlled vocabulary. I think this was really smart. And I think this is how we do, you know, movie reviews. It's basically a lot of people go in and say, thumb up, thumb down. And then The different journalists, they will look at this and look also at the movie and then say in which way this is a good movie or bad movie. And with all kinds of nuance, I think this was smart. And I think this is the way we should go. And I think everything in principle should be published, even the crappy stuff, but then associated with this crappy stuff, it says, you know, this don't trust this paper. Basically the conclusions are not substantiated by the data. Yeah. This is a lot better. But even at CoLab, actually, people don't agree with me. So, now at CoLab, they say, you know, you can't associate your brand name, Biophysics CoLab, with papers that are of too poor quality. So instead, we should reject them. And I think that's wrong. I think they should be, I think they should be in the open. I think we should just see them because otherwise, you're blind, like what do you yeah, I don't know. Maybe there's counterarguments to this, but in any case, it's not essential.

Milosz:

see always the edge cases where, you know, there's some academic war between two PIs and they downvote each other's papers or things like that. This is gonna happen at some point. But maybe there's a way to prevent that.

Lucie Delemotte:

I think you bring up a really good point, which is the fact that somehow humans will corrupt whatever system we design.

Milosz:

Yeah, that's it. I don't remember which law it is, but it basically says that whatever number is used for evaluation, it will

Lucie Delemotte:

Yeah, yeah, exactly.

Milosz:

or abused

Lucie Delemotte:

Yes. Yeah, exactly. It's, uh, it's when, uh, when something that used to evaluate becomes a target, it starts failing, but I forget what the exact sentence and citation is.

Milosz:

Maybe that's the problem.

Lucie Delemotte:

it's true.

Milosz:

now a a bit of context for our listeners. So you were supposed to be among the first guests of the podcast, when the podcast was kind of born, when I was staying at your lab a last winter. But then you gave birth to twins and you had a break in, uh, Sweden, which is famous for reasonably generous. leaves. what are your experiences? I mean, for the people out there who are planning to have kids, thinking about having kids in academia, is there something that surprised you really about this experience, you want to share with,

Lucie Delemotte:

Yeah.

Milosz:

with the folks

Lucie Delemotte:

Yeah, thanks for bringing it up. It's a big part of my life right now, so I'm happy to talk about it. and thanks for thinking of me for the podcast. I think it's very good that you were persistent in the, in the way you were and you insisted to do it and to invite me. I, I think this is a really fun,

Milosz:

right

Lucie Delemotte:

good resource. Yeah. Um, yeah, the twins were born now almost 15 months ago, and I was on leave actually for an entire year. And like you say, this is a Sweden makes this possible. But that's not the only point is also I think I was at a Career stage where I could do it and it goes back to all of what we've been talking about in terms of evaluation and so on. I didn't know I wanted children for quite a long time. I was actually kind of, scared about the world we live in and how they would be raised in this kind of world. So for a long time, I. Feels like, uh, I didn't really want children. I was also very satisfied, I think, with my job and my career and very busy. And then, uh, somehow it started, you know, they talk about biological clock. I guess that's what it is. But, then, uh, they came. And in fact, I think the fact that I was. in Sweden and in this kind of environment probably made a difference in my thinking about it, because I think other people will make very different choices based on really the societal conditions that they're living under. and I will say there are people, including women, even in Sweden, who don't make this choice to be away from the lab for that long. and I'm really glad I did it, just because I feel like these are probably my only children and they became two at once also, which I think is a different story, but, uh, maybe it's, uh, does increase a bit at least the amount of, work that you need to put into it. But, uh, yeah, the, I'm, I'm happy I did it. It was, it's, it was amazing. It was like, Taking on a completely new different job, basically, but it was equal. It was more demanding than the job. I thought, then being a, you know, assistant associate professor, which I thought was already kind of overwhelming in terms of the amount of, brain space that it takes. But it's a privilege also. And in fact, I think having children is very interesting for a scientist, just, you know, seeing them sort of evolve and watching them. But imagine

Milosz:

learning system.

Lucie Delemotte:

identical twins. It's like a dream. And it's a lot of fun in terms of, you know, imagining how they will evolve and become best friends, hopefully, or also best enemies. But from the scientific point of view, I'm totally fascinated. So they have, they really have the same DNA and yet they're different and they were different as soon as they came out in some way. And they're also very similar. So I mean, we get I need to go on a different podcast, I think, to talk about this, but, um, it's really,

Milosz:

experiment anyway.

Lucie Delemotte:

really fascinating to run your own sort of experiment. And I mean this in the best, you know, loving motherly way. Like, uh, I'm not actually submitting them to trauma, I hope, in order to support my experiment, but, um, I think it's really, it's really kind of fun to have this opportunity but yeah. I made the choice or the non choice of having children kind of late, and it worked out well for me. I've been in contact, uh, you know, as it goes, like you become close sometimes to people who go through similar life experience at the same time. So I've been, uh, in contact with people who've had Children around the same time in me, but have been a very different career stages, like younger women who are still at the end of their PhD or postdoc years, and it sounds really stressful for them. And I feel like it's really unfair because, you know, when you have to navigate this thing of having a newborn, figuring it out. And at the same time, you're worried about your next career step, and the competition that you have, I think is, is really difficult and unfair. And it's,, I'm sure it's not easy for men also, or, you know, people at different family circumstances, but I think for women it's, it's more difficult just because both the biological aspect of it, but also just the societal sort of pressures that exist. So,

Milosz:

I think definitely there's some period where it's unequal.

Lucie Delemotte:

but I must say that, um,

Milosz:

we, try.

Lucie Delemotte:

I was really lucky cause I didn't have any fertility issues. Uh, I think the type of choices that I made many people then really bite their fingers just because they end up being unable to, get pregnant or to, they have a very difficult fertility journey. So I was really lucky in that respect. And, I, that's why I would hesitate to recommend this to anyone because, uh, I think the risky

Milosz:

Well, for sure. For sure, the society and the medical side have a lot of work do there, right? To make this experience available to people independent of their biological luck, let's say.

Lucie Delemotte:

Yeah. Yeah. Yeah, And I mean there is support, but it's still a very difficult journey for many people and many women also young, but it gets harder as time goes by. So at the same time, yeah, I don't know. My life has been much improved now having these children in it. So yeah.

Milosz:

that's great.

Lucie Delemotte:

And I, it. I really think that, even though I love this job and I love the science, I feel so lucky that I was able to put it aside for a significant portion of time to be able to focus on this other thing, because everybody's a little bit different, but I find it quite hard to, divide your attention, that much. when there's pressure. So, you know, the first month are a lot of pressure just because these babies are just so dependent on you. and then if you're at a career stage, or you know, you're relying on the next grant to get your salary, it's also a lot of pressure. so those I feel like, at least for me, they, I was lucky that they were sort of coming at different times, basically. Now that they're older and that they are able to go to daycare, I'm really enjoying the fact that I can focus on work for a period of time in the day and then I can focus on them for another period of time. I think this is really nice and this will evolve as they become more independent. I think I can go back more to, you know, focusing more on, on work. So that's also really nice to be able to do. And in Sweden you get to do this while keeping somehow some time for sleep and hopefully other fun activities. So yeah, I'm, I feel

Milosz:

Sweden.

Lucie Delemotte:

lucky, very lucky.

Milosz:

I see, yeah, just a reminder that there are enriching things to life outside of science as well.

Lucie Delemotte:

Oh yeah. But, Yeah, we're we're really lucky when we have access to the the science part of things. I feel like, you know, some people have like a lot of spiritual experiences and things like that. I feel like all the science and academia for me. It's been a bit like, both a community and also a way to think about life that has really, been important and helped me. So and Yeah.

Milosz:

Mm hmm.

Lucie Delemotte:

know, it's even important for things like raising children, I think, to have this academic, kind of culture. You can try to do things in a more rational and data driven way, while keeping your emotional availability.

Milosz:

PEG students.

Lucie Delemotte:

Yeah, I don't. Yeah, I'll finish with this. I find now that between like 12 and 14 months with the twins, I found it so cool to look at the little scientists in them. And I think this is natural to any babies. It's really cool. You know, they will actually experiment like if I make this role on this surface, does it like fall in that direction or not? You know, it's really like

Milosz:

Yeah, what

Lucie Delemotte:

doing. they're they're

Milosz:

of reality?

Lucie Delemotte:

doing physics, uh, on a daily basis and, you know, trying to fit things in other things and like, I don't know, textures and it's really fun. So, I think we're all made to be scientists. Then we choose different ways, but, uh, we

Milosz:

Let's stick

Lucie Delemotte:

this

Milosz:

that. That's a

Lucie Delemotte:

innate way of, uh, wanting to experiment with the world and understand it.

Milosz:

that's a great place to finish. Thank you so much, Lucie Delemotte I'm happy you finally made it after a year, so thanks for the conversation and the insights.

Lucie Delemotte:

so much.

Milosz:

Have a great day. Thank you for listening. See you in the next episode of Phase Space Invaders.