0:00
It's from Mathias here, who is our Chief Information Officer for ChemSpeed Technologies, and he'll be taking you through his presentation called Automation for Self-Driving Labs.
0:11
So without any further ado, let's give a round of applause to Mathias.
0:15
All right, good afternoon, everyone.
0:31
My name is Mathias.
0:31
I'm Chief Information and Chief Technology Officer at ChemSpeed for about three and a half years.
0:41
I'm engineer by training.
0:47
I'm going to try to convince you something about chemistry now in like 15 minutes.
0:51
We came up with a relatively new presentation.
0:54
You are my first audience and that I would like to walk through.
1:01
And we're going to start with just a short overview about ChemSpeed and then a bit challenge that we would like to solve for you in the future, how we build up a modular library that can help you in your automation digitalisation journey.
1:24
And I'll give you 2 examples, like one will be our benchtop solution, so the smallest offering that we have.
1:31
And then one will be the biggest that we have, which is called FlexShuttle, which covers and everything like a complete a product development factory, so to say.
1:43
And I'll give you one example of how this ties in with artificial intelligence on machine learning, make, test, analyse, loop.
1:53
And then I'll also would like to invite you for then visit our booth and let's just across the hall and our open day this Friday.
2:06
This as an overview.
2:08
So ChemSpeed is now over 25 years on the market, and we have our headquarters quite close to Basel where more than 70% of our staff are chemists or engineers.
2:22
We just recently got acquired by Bruker and we are now heading their automation division.
2:32
We're globally operating, so as I said our headquarters in Switzerland, but we also have offices in Germany, the UK, the US and a joint office in Shenzhen and in China.
2:47
And now going into the challenge you see already here.
2:54
So the goal would be, or is of course, that through automation and digitalisation, you actually have time to think about your chemistry, about the design of experiments that you want to do and not have to ask the person on the left side have to note down everything in your lab notebooks, have to keep track of what is happening, but actually being present for the scientific part and not for, you know, pipetting liquids around.
3:27
And here you see, we see on the left side the main challenges.
3:34
So the manual data gathering, which is very cumbersome, of course, even if you have an ELM or a LIM system, it still is not very streamlined.
3:46
Then also the data of your experiments, they're often very scattered across different labs and there's a lack of standardisation.
3:56
And of course also then the difficulty to, because everything is so that we're so scattered, to extract actionable insights.
4:05
And the solution that we propose is having automated data collection, have everything connected.
4:13
So everything that you do in the lab has already built in structured data.
4:22
So you can easily follow standardised protocols, and you can use the data that you generate very easily to hand over to a machine learning or whatever data analytics tools that you want to use. And with the effective goal that you have more time to think.
4:40
Now what I mean, I probably don't have to explain you, but of course, if you have more time to spend on your really core task, we see that it's also an improvement in decision making.
4:55
You will have better experimental designs so that we'll also advance the chemical solution space that you can explore.
5:03
And then of course, we hope this will also help to increase the efficiency.
5:09
So to solve the challenges that we see coming that they can be solved by chemistry.
5:15
And how we do that is what is here.
5:18
We have built a 5 dimensional library where we can build or mimic all your workflows that you currently do manually or say manually or automated in the labs.
5:35
And we start with what you see here on the bottom with the first dimension.
5:40
So you choose with your deck setup, so what kind of reactors you need, what kind of shakers and so on.
5:46
And then we have a second dimension, what we call robotic tools.
5:51
So we have currently over 70 proprietary robotic tools that mimic all your handling activities.
5:57
That can be just moving plates around, but it would also be pipetting, it would be overhead cryometric dispensing of solids, it can be measuring a pH, whatever you need, what you currently would do manually.
6:13
And then we have on top of that a third dimension, third party modularity.
6:19
So we integrate all the very specific to your science that you do third party analytic equipment.
6:29
So that could be just a microwave reactor that is very specific, but it could also be some like an NMR or an IR, whatever you need.
6:39
And then of course, you have now built an automated set up, but you also need a software to drive it.
6:48
And this is then the 4th dimension.
6:50
We have two layers of software, Autosuite and Arksuite, and they have first of all the control of the module that you just configured in that way, but then also Arksuite has a data trunk, which then opens up all the APIs for machine learning for your ELN, for your LIM systems.
7:13
And now this is an example of having one automated system, but you can then use our fifth dimension.
7:23
This is going to want to show you in a couple of minutes, 5th dimension.
7:27
So the combination of multiple of these configured instruments to a whole product development factory.
7:37
This is what we call FlexShuttle. And we call this bottom up digitalization.
7:44
So you approach us with your workflow, and you work together with our workflow architects to come up with a workflow that helps you automate your science.
7:54
We then design the system for you, and you can then use it to execute your experiments in a high output fashion.
8:08
And of course, because we have Arksuite as a software layer overarching this, you can then connect everything and everything is feeding into the same database.
8:21
And so if you have one of our solutions, a third party equipment or even manual work streams, everything ties into Arksuite and everything is connected to your experiment.
8:34
And I give you 2 examples, as I already said, one is Crystal.
8:38
It’s also at our booth; it's our solid dispensing benchtop solution.
8:45
There are two different variants.
8:46
I'll quickly show you a video.
9:43
So we've seen here it's on a quite small footprint and it allows you to do sub-milligram dispenses of solids.
9:52
And then it also has an API.
9:53
So if you have as a first step, a solid dispensing step and then want to bring it over to a Tecan or Hamilton for a dilution step, this is can easily feed the data over to it.
10:04
I think this is also a very important aspect.
10:08
And then my next example is our biggest offering, so FlexShuttle, which boosts the end to end discovery.
10:22
And this is an example for especially for a life science pharma application.
10:26
So you see, we start on the top left with a synthesis loop.
10:32
And then we also have the complete sample management and preparation included.
10:38
We have purification and the chemical analysis step.
10:43
And then of course, to complete the loop.
10:46
So you can actually work on your DMTA cycle.
10:50
You also have integrated the by USA as the biological testing on the top right.
10:56
So you basically have a software interface in your office, and you can done experiments and everything that you would normally have to do and work with different teams can be done in one fully automated solution.
11:09
And this is an example workflow of how this would look like.
11:17
I will jump over this and go there, jump directly into a video of it.
11:23
You see we have a shuttle track, but it could also be a mobile robot.
11:27
We have both offerings, and you have the MTP plates that travel around from module to module.
11:33
Each module is very dedicated.
11:35
There's only one function here.
11:38
We see the sample prep stage.
11:46
I've included our SMOLE technology so you can encapsulate everything, then also included with all the reactions for the workup for the analytics.
12:40
This is now very specific, the life science.
12:43
I will show you quickly another example, more in the materials field, with a bit more colour.
13:01
So here it did the same, exactly the same concept, but you see a bit more happening because they're bigger samples than on an MTP plate.
13:56
All right.
13:57
And then now of course we have this offering of the of this full factory that you can get.
14:04
But of course this also has this Arksuite layer on top that gathers us all the data and collects it.
14:10
And this can then be used and have here an example of IBM.
14:13
So over our API, any third party software can also interface with this experimental setup.
14:21
You can use it here for example, in combination with IBM’s RoboRXN or any machine learning or in this case retro synthesis tool and send the instructions over to our automation and the robot will then execute it.
14:37
And while it's executing, play all the data that it's gathering back.
14:42
And of course, this is showing preparation and also the reactions, but it of course can also then feed directly into a HPLC or UPLC and use this data to drive the reaction.
15:07
All right, to summarise, because I was informed we're running out of time.
15:14
We see automation and digitalisation, they both tie in together.
15:18
We try to make in such a way that it frees up your time, so we have more time to think about your actual science and design your experiments.
15:26
And we have a booth outside, as I already mentioned.
15:28
Please visit us.
15:29
We also have an open day because we're quite close on Friday, and you can even use this QR code to register if you want to visit us and get a tour of our factory.
15:39
Thank you.