WEBVTT 00:00:00.000 --> 00:00:02.730 Welcome to the Deep Dive! Today, we're tackling 00:00:02.730 --> 00:00:05.910 a really pivotal question in pharmaceutical development. 00:00:06.669 --> 00:00:09.390 How are parallel synthesis and automated reactor 00:00:09.390 --> 00:00:12.550 systems changing the game? Basically, how fast 00:00:12.550 --> 00:00:15.570 can we create new medicines now? It's a critical 00:00:15.570 --> 00:00:18.010 area, definitely. So you, our listener, are going 00:00:18.010 --> 00:00:21.570 to get a clear understanding of how these advanced 00:00:21.570 --> 00:00:24.510 tools, and also things like small scale experiments, 00:00:24.750 --> 00:00:27.250 computer control. Statistical design, too. Right. 00:00:27.309 --> 00:00:29.250 And statistical design, how all that is driving 00:00:29.250 --> 00:00:31.170 efficiency, getting therapies to pay. patients 00:00:31.170 --> 00:00:33.390 faster. Yeah, and you might have noticed things 00:00:33.390 --> 00:00:36.409 like the special feature section, new technologies 00:00:36.409 --> 00:00:38.869 and process research. There's real interest here. 00:00:39.250 --> 00:00:42.909 Exactly. So our focus today is pretty sharp. 00:00:43.649 --> 00:00:45.929 How do these approaches actually optimize the 00:00:45.929 --> 00:00:48.130 development process and improve scalability? 00:00:48.289 --> 00:00:50.570 And ultimately shrink that timeline for new drugs. 00:00:50.829 --> 00:00:54.219 Okay. So where do we start? Parallel synthesis. 00:00:54.899 --> 00:00:57.399 Let's. It's fascinating how these technologies 00:00:57.399 --> 00:01:00.219 sort of converge. Parallel synthesis at its heart, 00:01:00.299 --> 00:01:01.840 it just means making lots of different compounds 00:01:01.840 --> 00:01:03.820 at the same time. Instead of one after the other. 00:01:04.159 --> 00:01:08.120 Precisely. Not that old linear way. Imagine generating 00:01:08.120 --> 00:01:10.700 whole libraries of molecules side by side. And 00:01:10.700 --> 00:01:12.620 then you combine that with automated reactors. 00:01:12.840 --> 00:01:15.099 Yes. Yes. And these are often small systems, 00:01:15.359 --> 00:01:17.719 our source has mentioned, like 10 to 100 milliliter 00:01:17.719 --> 00:01:21.549 scales. tiny, really. But they give you incredibly 00:01:21.549 --> 00:01:24.189 precise control. You can monitor lots of different 00:01:24.189 --> 00:01:26.989 reactions happening all at once. Okay, let's 00:01:26.989 --> 00:01:29.170 dig a bit deeper there. So moving from sequential 00:01:29.170 --> 00:01:32.849 to parallel, what's the immediate hit for someone 00:01:32.849 --> 00:01:34.870 trying to find a new drug? Well, the biggest 00:01:34.870 --> 00:01:37.709 immediate impact is speed, a dramatic acceleration, 00:01:37.870 --> 00:01:39.790 especially in those early discovery stages. Right. 00:01:40.129 --> 00:01:42.450 You can suddenly screen a much, much broader 00:01:42.450 --> 00:01:45.150 range of conditions. Try different catalysts, 00:01:45.250 --> 00:01:47.530 temperatures, solvents. All at once. All at once. 00:01:47.769 --> 00:01:50.450 So the scope of your experimentation just explodes. 00:01:51.109 --> 00:01:53.209 And you get a richer understanding of the chemistry 00:01:53.209 --> 00:01:56.370 involved in way less time. So it's not just saving 00:01:56.370 --> 00:01:59.250 lab hours. Yeah. potentially changing what's 00:01:59.250 --> 00:02:02.290 even possible. Absolutely. It can alter the economics, 00:02:02.950 --> 00:02:05.150 make research that was maybe too slow or expensive 00:02:05.150 --> 00:02:08.509 suddenly viable. That makes sense. Now, we know 00:02:08.509 --> 00:02:11.750 early drug development often throws up some tricky 00:02:11.750 --> 00:02:14.889 challenges, doesn't it? Always. Like one source 00:02:14.889 --> 00:02:17.590 mentioned, poor water solubility. Lots of promising 00:02:17.590 --> 00:02:20.409 candidates have that issue. How do these parallel 00:02:20.409 --> 00:02:22.770 techniques help there? That's a great example 00:02:22.770 --> 00:02:26.449 because you can quickly make and test lots of 00:02:26.449 --> 00:02:29.689 variations analogs of your lead compound. Tweaking 00:02:29.689 --> 00:02:32.490 the structure slightly each time. Exactly. You 00:02:32.490 --> 00:02:35.330 systematically explore changes designed specifically 00:02:35.330 --> 00:02:38.129 to improve properties like solubility. Make a 00:02:38.129 --> 00:02:40.590 small change, test it quickly, see the impact. 00:02:40.909 --> 00:02:43.550 You navigate that chemical space much more efficiently. 00:02:44.189 --> 00:02:47.370 So it's like this engine for exploring possibilities 00:02:47.370 --> 00:02:49.819 around a starting point. trying to find that 00:02:49.819 --> 00:02:51.699 sweet spot, the right properties for a drug. 00:02:51.819 --> 00:02:53.939 That's a good way to put it. An efficient engine 00:02:53.939 --> 00:02:56.080 for molecular exploration. And you mentioned 00:02:56.080 --> 00:02:58.000 working small scale. Why is that so important, 00:02:58.180 --> 00:03:01.020 the miniaturization? Oh, it's huge for efficiency. 00:03:01.740 --> 00:03:03.919 Resource efficiency. You need way less starting 00:03:03.919 --> 00:03:06.439 material. You generate minimal waste. Which lets 00:03:06.439 --> 00:03:08.800 you do more experiments. Far more, yeah. Yeah. 00:03:09.080 --> 00:03:11.710 And critically, the data you get. from these 00:03:11.710 --> 00:03:14.669 small parallel reactors, it can actually be quite 00:03:14.669 --> 00:03:16.889 representative of what happens on a larger scale. 00:03:16.969 --> 00:03:19.210 Really? Yeah, one of the sources really emphasizes 00:03:19.210 --> 00:03:22.949 that. So you gain crucial insights without the 00:03:22.949 --> 00:03:26.389 massive cost and material needs of big trials. 00:03:27.169 --> 00:03:29.430 Interesting. So insights from tiny experiments 00:03:29.430 --> 00:03:33.889 can actually guide the scale -up process. Are 00:03:33.889 --> 00:03:37.979 there limits? pitfalls maybe? That's a very important 00:03:37.979 --> 00:03:40.620 question. It's not always a perfect one -to -one 00:03:40.620 --> 00:03:43.020 translation of course. Okay. Things like say 00:03:43.020 --> 00:03:45.460 heat transfer or mixing behave differently in 00:03:45.460 --> 00:03:47.680 a huge vat versus a tiny vial. That's physics. 00:03:47.819 --> 00:03:50.860 Right. But having that fundamental understanding 00:03:50.860 --> 00:03:53.719 early on optimal conditions, potential impurities, 00:03:54.180 --> 00:03:56.340 reaction kinetics, it's incredibly valuable. 00:03:56.439 --> 00:03:58.300 It means you make much more informed decisions 00:03:58.300 --> 00:04:00.680 when you do scale up. So you avoid major surprises 00:04:00.680 --> 00:04:03.860 later, less costly rework. Exactly. You reduce 00:04:03.860 --> 00:04:05.979 the likelihood of hitting big problems. Think 00:04:05.979 --> 00:04:08.780 about that example from one of the sources. Synthesizing 00:04:08.780 --> 00:04:11.340 over 500 -quinoxyline derivatives. For anti -TB 00:04:11.340 --> 00:04:13.719 drugs, yeah. Right. You just couldn't explore 00:04:13.719 --> 00:04:17.100 that much chemical space practically or affordably 00:04:17.100 --> 00:04:20.139 without doing it small scale in parallel. 500, 00:04:20.279 --> 00:04:23.189 wow. That really hammers home the point. And 00:04:23.189 --> 00:04:27.089 didn't that same source mention modifying exocyclic 00:04:27.089 --> 00:04:29.889 groups? Yes, for fighting drug -resistant TB. 00:04:30.129 --> 00:04:32.589 Where even tiny structural changes had big effects 00:04:32.589 --> 00:04:34.889 on activity. Precisely. Like the position of 00:04:34.889 --> 00:04:37.370 one methyl group. That level of fine -tuning 00:04:37.370 --> 00:04:40.350 is perfect for these automated parallel setups. 00:04:40.629 --> 00:04:43.180 You make the tweak, you test it fast. A very 00:04:43.180 --> 00:04:45.920 efficient way to optimize how the molecule hits 00:04:45.920 --> 00:04:48.240 its target. You got it. OK, let's shift to the 00:04:48.240 --> 00:04:50.639 computers. The control and automation side. It's 00:04:50.639 --> 00:04:52.860 not just doing many things at once. It's the 00:04:52.860 --> 00:04:55.060 way they're done, right? The precision. Absolutely 00:04:55.060 --> 00:04:57.939 critical. Computer control gives you incredibly 00:04:57.939 --> 00:05:00.480 accurate handling of all the key parameters. 00:05:00.939 --> 00:05:03.509 Temperature. stirring speed, how fast you add 00:05:03.509 --> 00:05:06.569 reagents. Very tight control. Very tight. Which 00:05:06.569 --> 00:05:09.629 leads to much more consistent reproducible experiments 00:05:09.629 --> 00:05:12.129 compared to, well, doing it all by hand. And 00:05:12.129 --> 00:05:14.850 that consistency must be vital for trusting the 00:05:14.850 --> 00:05:17.689 data you get. Oh, absolutely. Plus, these systems 00:05:17.689 --> 00:05:20.490 usually log data automatically. They monitor 00:05:20.490 --> 00:05:24.009 progress, yield, impurities, generating this 00:05:24.009 --> 00:05:26.089 huge amount of data that's instantly stored and 00:05:26.089 --> 00:05:28.629 ready for analysis. Much easier than scribbling 00:05:28.629 --> 00:05:31.089 in a lab notebook and trying to spot trends later. 00:05:31.259 --> 00:05:35.759 So speed, efficiency, tons of high quality, consistent 00:05:35.759 --> 00:05:39.019 data. But how do you decide which experiments 00:05:39.019 --> 00:05:41.480 to run? Out of all the possibilities, that's 00:05:41.480 --> 00:05:44.459 where statistical design comes in, DOE. Yes, 00:05:44.699 --> 00:05:47.459 you've hit on a really key piece. Design of experiments, 00:05:47.639 --> 00:05:50.139 DOE. It gives you a smart strategy for planning 00:05:50.139 --> 00:05:52.519 these parallel runs. So not just random trial 00:05:52.519 --> 00:05:55.819 and error. No, no. It's systematic. DOE lets 00:05:55.819 --> 00:05:58.800 you very... multiple factors, temperature, concentration, 00:05:59.019 --> 00:06:01.459 time, whatever, all at once, but according to 00:06:01.459 --> 00:06:03.759 a specific efficient plan. Okay, so instead of 00:06:03.759 --> 00:06:05.560 changing just one thing, seeing what happens, 00:06:05.600 --> 00:06:08.439 changing another, you adjust several things together, 00:06:08.560 --> 00:06:10.899 systematically. What's the big win there? The 00:06:10.899 --> 00:06:14.199 win is efficiency and insight. By varying factors 00:06:14.199 --> 00:06:16.519 together, DOE helps you pinpoint the critical 00:06:16.519 --> 00:06:18.439 parameters, the ones that really drive the reaction 00:06:18.439 --> 00:06:20.740 outcome. The important levers to pull. Right. 00:06:21.060 --> 00:06:23.360 And crucially, how those parameters interact 00:06:23.360 --> 00:06:27.019 with each other. Which you often miss with a 00:06:27.019 --> 00:06:29.920 one factor at a time approach interactions. Okay, 00:06:30.019 --> 00:06:32.860 so you get a much deeper more comprehensive understanding 00:06:32.860 --> 00:06:37.050 of the process but with significantly fewer actual 00:06:37.050 --> 00:06:39.870 experiments. It adds statistical rigor. And that 00:06:39.870 --> 00:06:42.490 rigor helps build a robust process, something 00:06:42.490 --> 00:06:45.430 reliable. Exactly. Because robustness is what 00:06:45.430 --> 00:06:47.269 you need for manufacturing, isn't it? Yeah. You 00:06:47.269 --> 00:06:49.470 want a process that works consistently, even 00:06:49.470 --> 00:06:51.389 with the small variations you inevitably get 00:06:51.389 --> 00:06:54.250 in a real plant. Makes sense. So DOE helps you 00:06:54.250 --> 00:06:56.769 find not just the best conditions, highest yield, 00:06:57.209 --> 00:07:00.050 best purity, but also the acceptable range around 00:07:00.050 --> 00:07:02.509 those conditions, the operating window. So you 00:07:02.509 --> 00:07:04.389 know how much wiggle room you have. Precisely. 00:07:04.459 --> 00:07:06.800 And it also helps you understand and quantify 00:07:06.800 --> 00:07:09.819 variability in the process itself, which is key 00:07:09.819 --> 00:07:12.379 for ensuring consistent quality batch after batch. 00:07:13.040 --> 00:07:15.279 It connects back to needing consistency for reliable 00:07:15.279 --> 00:07:18.079 results, like in clinical trials. A well -understood 00:07:18.079 --> 00:07:21.259 process gives predictable outcomes. OK, so let's 00:07:21.259 --> 00:07:24.079 tie it all together. We have parallel synthesis, 00:07:24.680 --> 00:07:26.939 automated reactors, small -scale work, computer 00:07:26.939 --> 00:07:30.939 control, statistical design. How does this whole 00:07:30.939 --> 00:07:33.420 package speed up getting a drug from the lab 00:07:33.420 --> 00:07:36.980 bench to, well, to patients? Well, all that knowledge 00:07:36.980 --> 00:07:40.100 you gain early on from those small, controlled, 00:07:40.300 --> 00:07:43.339 statistically designed experiments, it directly 00:07:43.339 --> 00:07:45.980 fuels process optimization. You figure out the 00:07:45.980 --> 00:07:49.139 best way to make it much earlier. Yes. You identify 00:07:49.139 --> 00:07:51.480 efficient, robust conditions right at the start. 00:07:51.720 --> 00:07:53.839 So when it's time to scale up for manufacturing. 00:07:54.100 --> 00:07:56.279 Fewer nasty surprises. Less troubleshooting. 00:07:56.500 --> 00:07:59.000 Much less, ideally. Early optimization heads 00:07:59.000 --> 00:08:00.879 off many problems that might only show up at 00:08:00.879 --> 00:08:03.120 large scale. Some sources talk about the challenges 00:08:03.120 --> 00:08:05.480 of scale up, and this approach really tackles 00:08:05.480 --> 00:08:08.060 that proactively. And fundamentally, just being 00:08:08.060 --> 00:08:10.000 able to identify those optimal conditions so 00:08:10.000 --> 00:08:12.079 much faster drastically cuts down the overall 00:08:12.079 --> 00:08:14.180 development timeline. Moving from discovery to 00:08:14.180 --> 00:08:15.819 clinical trials can happen much more quickly. 00:08:16.139 --> 00:08:18.459 The small -scale insights really do inform and 00:08:18.459 --> 00:08:20.699 streamline those bigger pilot plant runs. OK, 00:08:20.740 --> 00:08:23.839 so to sum it up then, key takeaways on how these 00:08:23.839 --> 00:08:26.500 technologies accelerate development. I'd say, 00:08:26.720 --> 00:08:28.959 number one, a seriously compressed timeline, 00:08:29.220 --> 00:08:31.079 because you explore and optimize so rapidly, 00:08:31.540 --> 00:08:34.080 two, more efficient process optimization, leading 00:08:34.080 --> 00:08:37.120 to robust, reliable manufacturing methods, and 00:08:37.120 --> 00:08:41.200 three, improved scalability. That transition 00:08:41.200 --> 00:08:43.580 from lab bench to factory floor becomes much 00:08:43.580 --> 00:08:46.190 smoother, much faster. Ultimately, it sounds 00:08:46.190 --> 00:08:48.610 like these tools are directly tackling that urgent 00:08:48.610 --> 00:08:50.450 need, getting new treatments out there sooner. 00:08:51.009 --> 00:08:53.049 Absolutely. And with innovation constantly happening, 00:08:53.110 --> 00:08:54.889 like those new technology sections highlight, 00:08:55.230 --> 00:08:57.029 we can only expect these tools to get better 00:08:57.029 --> 00:09:00.190 and have an even bigger impact. So as you, our 00:09:00.190 --> 00:09:02.370 listener, think about all this, it really makes 00:09:02.370 --> 00:09:04.750 you wonder, doesn't it, how will pharma development 00:09:04.750 --> 00:09:07.110 keep evolving as these technologies get even 00:09:07.110 --> 00:09:09.750 smarter, even more widespread? What else might 00:09:09.750 --> 00:09:12.289 they unlock down the road? It's definitely a 00:09:12.289 --> 00:09:14.350 space to watch. Thanks for joining us for this 00:09:14.350 --> 00:09:14.870 deep dive.