Chromatography is a powerful a flexible technique at all stages of chemistry research. Whether you are a graduate student attempting purify a compound or a pharmaceutical company attempting to test product quality, an effective chromatographic method is critical.

But how do you develop these methods? Optimizing chromatography techniques is notoriously time consuming and frustrating work. Can predicative software increase the productivity of chromatographers?

Learn more about chromatography software from ACD/Labs: AutoChrom.

Read the full transcript

Justin Shearer   00:00

It allows the opportunity for you to democratize chromatography method development. You don’t need to have a Ph.D. in chromatography or have done it for 25 years to develop completely adequate and acceptable methods. But that’s a double-edged sword. By democratizing that, anybody can do it. But what happens when it falls over?

Jesse Harris   00:30

Chromatography is a powerful analytical tool. It is one of the best ways to learn about the composition of complex mixtures.

Charis Lam   00:38

But that doesn’t mean it’s easy. Developing a chromatographic method is notoriously time-consuming and frustrating

Jesse Harris   00:44

Hi, I’m Jesse.

Charis Lam   00:46

I’m Charis. We’re the hosts of the analytical wavelength brought to you by ACD/Labs.

Jesse Harris   00:52

In this episode, we’re going to talk to a scientist from GSK about the advantages of computer assisted chromatographic method development in a large pharmaceutical organization.

Charis Lam   01:02

But first, we’re going to talk with Dr. Scott Van Bramer. He’s a professor at Widener University, and he’ll provide an introduction to using predictive tools in chromatographic method development by describing his experience teaching the software to his students.

Jesse Harris   01:18

Joining me on the podcast today, I have Scott Van Bramer. He is a professor at Widener University. Thank you so much for joining us today, Scott.

Scott Van Bramer   01:28

Thanks, Jesse. It’s a pleasure to be here.

Jesse Harris   01:29

It’s a pleasure having you. So to begin with, we want to, of course, start with our icebreaker question and that is, what is your favorite chemical?

Scott Van Bramer   01:39

It’s always hard to narrow questions like this down to just one answer. So I’ll give you a one and then maybe I’ll add a second one in as a bonus round.

Jesse Harris   01:47

I’ll allow it.

Scott Van Bramer   01:49

So the first one I like to answer is a pamoic acid, which has a really, really interesting structure, and it turns out it’s the best NMR spectroscopy unknown compound I’ve ever found. And so I like it. It’s also a really bright, pretty yellow. So that’s my first answer. And then my second answer would probably be ethanol, which is just a great molecule to use in freshman chemistry, as a structure both to get students attention, and because when we’re just first starting to talk about the idea of atoms arranged in molecules in different ways, it’s just a really easy way to introduce that. So those are my two favorites. Plus, I like having a good beer after work, so.

Jesse Harris   02:48

I was going to say you can’t go wrong with ethanol. You know, that’s a crowd favorite. So yeah, it’s a choice. Good. Okay, so let’s get into things. Can you tell me a little bit about your experience with teaching method development, particularly before the pandemic? Because I understand that you’ve been involved in teaching analytical method development for some time now.

Scott Van Bramer   03:08

Yeah. So I teach an undergraduate course on instrumental analysis, and one of the things that I really want students to get out of there is shifting from this idea that instrumentation in measuring is just a magic black box where you open up a drawer and you put something in and then the answer comes out. And so all of that, for my whole career teaching that course, a lot of the emphasis has been on method development—going in, running an experiment, and then changing something to see if the results get better; if the measurement gets more sensitive, if the noise goes down, if it goes faster, whatever. And so that’s always been one of the major things that I try to teach in that course.

Jesse Harris   03:56

Okay. And then a lot of that, though, of course, involves working with instrumentation. So I understand that that got a little bit more difficult during the pandemic, particularly in the early stages when a lot of university labs were essentially shut down.

Scott Van Bramer   04:15

Yeah. And so the timing couldn’t have been better. I was teaching instrumental analysis that semester, and we were just gearing up to do the first HPLC chromatography experiment. We were going to look at the caffeine content in student’s favorite beverages. And so that that could have been my favorite molecule. And we were actually doing a brainstorming session the whole class. We were all together talking about method development and students were trying to figure out what solvents to use and what HPLC column we should use and things like that. When one of the students got distracted by the text alert from the university saying that everybody had to leave right away and we were going to go into lockdown.

So I went home that night and drank, but then I remembered a few years before I had played around with some ACD[/Labs] software and I knew that they had an HPLC sort of method development suite. And so I reached out to the salespeople there, and that’s how I came up with my solution.

Jesse Harris   05:37

Good. And then how did your students incorporate the method development software in their lab work? So like what did that kind of like look like practically, when they were working with this software as part of their practice.

Scott Van Bramer   05:51

So what we ended up doing was based on that initial work that we’d done to sort of come up with an initial method, the students were aware of sort of what parameters might be important to change in class and so the way this was set up is I had one license and a computer that could run the software. And so then what I had was the students were working in groups for that experiment. And so I had each group come in and on Zoom, I shared my screen and then asked them questions about what they wanted to change. And so then we would just use the HPLC method software and we would change the flow rate and see what kind of an impact that would have on the simulated chromatogram. And so we were able to use the software as the focus of a discussion where we explored possible method parameters.

Jesse Harris   06:46

And one of the like the themes, I think of this episode and then our season in general, is about the danger of predictions and these don’t just replace more traditional ways of working or teaching. How does software help in teaching method development and where is it not as useful? So what is the, I guess, the areas in which it helps and the areas in which it’s maybe not there yet?

Scott Van Bramer   07:14

So for the first time that I did it, the way it helped was that we could do it on the computer on Zoom during the pandemic where nobody could be on campus What I’m looking at, so I’m teaching the class again for the first time again this semester. And what I have in mind is to use the method development software to help speed things up. So one of the things I found when we were playing around with this, is that the simulations were faster than the experiments. And so instead of running an HPLC experiment for 15 or 20 minutes, and then changing the flow rate, and then running it, letting it equilibrate, and then do it again; or, you know, trying a couple of different mobile phase mixtures and then having the conversation about switching to a gradient which could easily use a whole three hour class period, laboratory period. My plan is to spend a little time investigating that space, and seeing what they’re going to be… so that they’ll know what they’re looking for. And then during that three hour lab time that we have, we can use that time a lot more efficiently.

Jesse Harris   08:38

Yeah. And I think that also you were mentioning earlier about this idea of treating analytical methods as a black box that you just plug and play into. And I imagine that that’s also something that comes up in using this method development software as well; of like understand… using this as a way to not just say, you know, spit out a method, but understanding why it’s coming out to the results that it gives you and using a particular chromatographic method.

Scott Van Bramer   09:08

Yeah. And I’m going to be really curious to see sort of how the students respond to that. And how much they ask questions about like how’s it working? What’s going on behind the scenes? Because, you know, when I when I used this last time, I was using the demo templates that came with it to play around with because I couldn’t collect data for the training sets. And so, you know, I haven’t completely thought through how we’re going to use it; if we’re going to try and come up with our own training sets and then use that for the optimization. I got to spend some more time with the software and some more time with the new instrumentation that we got that I think is going to be more compatible with it.

Jesse Harris   09:54

Good. Okay. This is the last question that I want to ask is what advice would you have for professors or instructors who are trying to incorporate software simulation into their analytical chemistry labs?

Scott Van Bramer   10:07

My thoughts on this, in my experience, sort of what’s useful is that the software can give students a place to sort of sandbox, a place to play around, to twist and turn knobs or change variables and see how that could impact things. But they still need to make sure when they go down to the lab that like the instrument’s plugged in and there’s a level of real life troubleshooting that I don’t think that the software can replace. Even software that simulates noise and tries to simulate reality; it’s never going to be reality.

Jesse Harris   10:49

Well, I mean, there’s a certain element of just, you know, managing HPLC. It’s something that, in my graduate work, I did a lot of work with HPLCs, and they can be temperamental machines.

Scott Van Bramer   10:59


Jesse Harris   11:02

So nothing can quite replace that, I feel. But anything else that you want to add before we wrap up?

Scott Van Bramer   11:07

No, just, you know, I want to thank ACD[/Labs] software for providing me with access to this program during the pandemic—during this sort of emergency pivot—and it’s been really helpful to me and I think it’s improved and been a valuable experience for my students.

Jesse Harris   11:26

Great. Okay. Well, thank you very much for your time then. And take care.

Scott Van Bramer   11:30

Thank you, Jesse. Be well.

Jesse Harris   11:33

This was a great starting point for our discussion. The idea of a sandbox to explore analytical methods is a useful analogy for chromatography software.

Charis Lam   11:44

Now let’s take things to the next level. What does it look like when the software is used in a large company?

Jesse Harris   11:49

Dr. Justin Shearer is a separation scientist and team leader at GSK. Let’s talk with him about how computer assisted chromatography tools impacted his work. Joining me today, I have Justin Shearer. He is a separation scientist and team leader at GSK. How are you doing today, Justin?

Justin Shearer   12:09

I’m doing really well today, Jesse. Thanks for having me. Really happy to be here and see how this goes.

Jesse Harris   12:15

I’m very happy to have you as well. So let us start off with our icebreaker question. What is your favorite chemical?

Justin Shearer   12:22

You know, I’d like to be witty and come up with something absolutely off the wall. But, you know, if it weren’t for caffeine, I wouldn’t get to do half the things I do. So obviously, like most of us, big caffeine addict and happy to claim that.

Jesse Harris   12:34

Good, good stuff. Well, we were mentioning just before we recorded that you had an early morning this morning too, because GSK is, of course, between both the US and the U.K., so I imagine there are some early mornings there.

Justin Shearer   12:44

More than I’d like to admit.

Jesse Harris   12:50

Good stuff. Okay, so let’s get into things. How did you get involved in computer assisted chromatography method development?

Justin Shearer   13:00

Yeah, my start with a computer assisted chromatography method development actually started in a previous company. And at that point in time I was in the lab doing a lot of small molecule method development and doing that by brute force—manual method development where I’d select columns and work my way through different columns, different mobile phase combinations, and ultimately relying on my own brain and training to arrive at the best conditions.

And I got into a project where I was just having a lot of trouble and started having some conversations with some colleagues across that company and asked what they knew about where we were using this in the company and using chromatography method development software. And lo and behold, we actually had a license on site that nobody was using for a specific piece of software and I jumped in feet first, realizing that, hey, this is probably an opportunity to make my life a whole lot easier.

I spent a couple of weeks scouting columns, spent a couple of weeks optimizing mobile phases, and very quickly, within about three weeks, I was able to spin up modifying an HPLC stack to include a single quad to allow me to do some peak tracking as well as a diode array detector, getting an eight column compartment in to allow me to actively test different columns. And during the training that came along with that license, I actually used the project that I was working on. And during that one week of training, I was able to develop somewhat adequate conditions. And I’d already been at this for about a month. So really happy that in the course of a week of training, I was able to derive conditions that I was able to progress to robustness.

And that to me was my start. And immediately I was sold on the actual utility of In Silico Chromatography Method Development software.

Jesse Harris   14:51

Yeah. And from your description there, I mean, I, in the course of my studies, did some method development stuff for HPLC, and it can be incredibly time intensive, you know, for doing this. But I think that there’s also a question of how these compare—of the experimental based versus computer-based approaches. Do you see computer-based methods as just accelerating the same work or is it fundamentally different in some capacity?

Justin Shearer   15:19

Yeah, I think that’s a very interesting question, Jesse. What I have actually seen is and I’ve actually employed multiple formats, you know, as I mentioned, I started using brute force chromatography development. I now call that ‘any column will do’—you basically pick a column off the shelf and you try it. And if it doesn’t work, you pick another column. You know, sometimes you’re doing it without any knowledge of the chemistry that you’re actually looking to separate, or any of those impurities and therefore, it can be a very time-consuming process. I’ve also worked in things where in groups and have implemented processes where we have used platform columns and we’ve maybe, you know, had six columns in a couple different mobile phase combinations and then done a lot of manual optimization around those columns for things coming out of the laboratory.

And also worked in a system where, here at GSK in particular, we actually have a suite of columns that we leverage and then we have prescribed gradients and sometimes people use in silico software, sometimes they don’t, you know, depending on what comes out of those gradients, and all of those are actually useful in time and can be time consuming. And as you mentioned, you know, the use of the In Silico Software can expedite that process of coming up with those more efficient conditions and really, I have enjoyed that opportunity to (whether you’re doing a full control or not) basically interpolate around some boundaries. And that’s where most people actually get engaged with chromatography method development software is collecting a shallow gradient and a long and a steep gradient; so they have long chromatography conditions, short chromatography conditions, and they’re basically tracing retention factor on specific compounds and creating boundary conditions and then interpolating between those to come up with somewhat optimized conditions. That does give you a way to be more optimized. but to me it can be a little bit more fundamentally different in that it allows the opportunity for you to basically democratizing chromatography method development. You don’t need to have a Ph.D. in chromatography or have done it for 25 years to develop completely adequate and acceptable methods. But that’s a double-edged sword, right? By democratizing that, anybody can do it. But what happens when it falls over, is ultimately the question that comes about. And so without that trained expertise or some inherent knowledge of what separations actually do, you can really get yourself into some deep water pretty quickly and not be very successful.

Jesse Harris   17:58

That’s an interesting to hear. And I think that that might relate to the next question I want to touch on. GSK is obviously a large pharmaceutical company that has several labs spread across multiple countries, not to mention any outside companies that you might be partnering with. Do you feel that computer-based approaches to method development help your collaboration?

Justin Shearer   18:20

They obviously do. Anything we can do to make what we do more consistent in these different geographies, to ensure that we’re using the same columns or the same mobile phases, is actually just in parts a real opportunity for us to be more efficient and effective as we take our candidate molecules from our early phase, pre candidate selection all the way through to manufacturing.

If we can get to the point where we know that our manufacturing, whether it’s internal or external, only has to maintain a certain type of instrument; the HPLC stack itself is only potentially going to use one of six to ten different columns and use the same mobile phases. We actually are driving predictability to what we’re leveraging in that longer term strategy in the lifecycle of our molecules as we’re going into patients; as long as we are focusing on that understanding the science and the impurities and the real opportunities for us to make sure that we are providing high quality drugs and medicines to our patients, we can be consistent and really drive efficiency as well as consistency, which is really good.

One of the things that we haven’t quite touched on yet is one of the opportunities that chromatography software allows you to do is it allows you to focus on getting to really strong conditions that you can then really focus on the robustness and ruggedness and make sure that what you are transferring out to those external manufacturing, or internal manufacturing partners, is going to work day in, day out, and not impede the release of that transformational medicine so people can do more, feel better, live longer. Notice I worked the catch phrase in.

Jesse Harris   20:11

Yeah, and I think that people should appreciate the fact that methods might be developed in a method development specialist, but they are actually being used there more on the factory floor basically, and that’s not exactly the same.

Justin Shearer   20:26

Yeah. You know, that’s one of the most interesting things that you think about whenever you first join in industry and you work in R&D in particular, you think ‘What I do stays here.’ And then you realize very quickly that, oh, no, this goes out to the whole world. It doesn’t matter if it’s pharmaceutical, in the specialty chemicals industry, or even in agricultural chemistry. There’s, you know, any type of regulated industry, over-the-counter medicines, what you’re doing impacts people at the end of the day. And there has to be this opportunity to think globally about what you do, but acting locally on developing for your projects and delivering those assets to the best of your ability.

Jesse Harris   21:02

Yeah, I mean, that’s a whole interesting thought that just thinking back to my time in doing research in in university, because it’s something that you know, the methods that you’re using are really going to be used by you and maybe a couple of other students, but here it just has such a bigger implications. But that’s maybe a conversation for another time. With that, though, I wanted to talk about what happens after this is done; once the computer work is done. Is there more work to do afterwards or is it kind of, okay, we have the computer method developed for us, let’s just plug and play?

Justin Shearer   21:32

So once we actually have those in silico conditions, we do have to go back and derive empirical data on that. And so we have to demonstrate that the method that the computer tells us, that the modeling tells us, does work for the compound of interest, and does provide adequate resolution and the necessary information to ensure adequate quality for patients. So once we have those conditions, what we actually need to do is test them, and then we will often go through a ruggedness and robustness assessment where we make sure that we understand the boundaries of that method that comes out. A lot of times we’re not happy if we’re walking on a knife’s edge. That’s not going to be something that’s going to translate and transfer to our manufacturing partners and allow lifecycle to continue for that product. We’re happy for working on wide plateaus. You know, we want to have broad areas that if we happen to make the mobile phase incorrectly or the temperature of the column compartment isn’t necessarily what we set it to. And we want to make sure that the method actually works around some range of conditions and once we have that ruggedness and robustness assessment, we will still go through full validation criteria and then transfer of that methodology to receiving laboratories, whether it’s a pilot plant that’s doing some scale up studies or it’s going out to our full manufacturing partners, we still will transfer that and we will often still continue to support that method even after it goes out of our shop and we may not be working on it because that molecule is being produced for delivery to patients, we will still support that if some trouble comes up along the way. Unfortunately, we never actually step away from it, even though we see less of it.

Jesse Harris   23:18

Yeah, no, that makes sense that no method ever truly, truly goes away. It leaves the building but you are kind of responsible for all of them still. And I think that that also ties into this next question as well. Do you have any other thoughts or advice for scientists who want to get started in automated method development? I think there’s a lot in what we’ve discussed already, but was there anything that you wanted to add?

Justin Shearer   23:45

So a couple other things I like to say that are other thoughts and additional thoughts for scientists who are getting into this. The thing I would like to say is stay tuned. This is an ever evolving field at the moment. We have seen massive growth in the use of chromatography method development, particularly in the small molecule space. Most of us in pharmaceuticals and other industries are using it for reverse phase, normal phase, chiral separations, and even a couple of others. We’re seeing opportunities to create databases and do a pseudo QSRR approach where we might have a new compound come in and we may put that in against our library and say, oh, this column in these conditions worked for this compound, which looks very similar. A lot of growth in that space, really looking forward to seeing how we continue to grow in full blown AI/ML in the QSRR space that’s taken a lot of shape over the past ten years, and there’s a lot of growth to be had in that space, particularly when it comes to thinking about gradients, etc.

But other areas where I’d really like to see a lot of growth is in the large molecule space, in particular biopharmaceuticals, monoclonal antibodies, oligonucleotides, the medium sized molecules that we’re starting to work with across the industries as well as mRNAs. A lot of recent work has been done in these larger molecules, from the 10,000 molecular weight, to the 150,000 molecular weight, up to the mega dalton mRNAs. A lot of information coming in that way.

Chromatography method development is at its core ultimately a good mix of you knowing something about your compound, having the right column in the right mobile phase to actually tease things apart so that you can discern small differences in impurities to ensure efficacy, safety for whatever that intended impact is. So number one, it’s about the column, it’s about the chemistry. Without that you fall over. But where chromatography development can really be helpful and beneficial is maybe enabling you to arrive at more robust conditions that then you can focus on the robustness because you know more information about your system than just focusing on getting to adequate conditions. So maybe it’ll improve what we do in terms of method transfer and allowing us to have high quality method.

My hope is that it goes for not just small molecule but every mode of chromatography that is important for release testing, for whatever that application is.

Jesse Harris   26:19

Wonderful. Thank you so much for joining me. That was a great overview of the subject it’s been a pleasure having you.

Justin Shearer   26:27

Thanks a lot, Jesse. I hope it was helpful and thanks.

Charis Lam   26:31

We keep coming back to that idea of using predictive tools responsibly. It takes real chemical expertise to both use these tools effectively and implement them.

Jesse Harris   26:42

That is going to be especially important in our next episode. We are closing out our season with a look into A.I. assisted drug discovery. Is it worth the hype?

Charis Lam   26:53

Be sure to subscribe so you don’t miss our interview with Dr. Chris Lipinski. Famous for developing the original prediction tool, the Lipinski Rule of Five.

Jesse Harris   27:02

This has been the Analytical Wavelength, see you next time!

Charis Lam   27:06

The Analytical Wavelength is brought to you by ACD/Labs. We create software to help scientists make the most of the analytical data by predicting molecular properties, and by organizing and analyzing the experimental results. To learn more, please visit us at

Enjoying the show?

Suscribe to the podcast using your favourite service.