Human Practices

Here goes more text

Introduction

We are still updating our wiki, stay tuned!

Interviews with female scientists

Laura Wiens

PhD student
Laura is a 28 years old PhD student in Prof Susanne Gebhard's research group for two years now. She started studying molecular biology in 2016, then joined Prof. Dr. Ralf Heermann's research group, wrote her Master's thesis there and then joined Susanne Gebhard's microbiology group.

Laura Wiens

Was it immediately clear to you that you wanted to become a scientist?
Not really. I always had good grades in written work, but otherwise I was always quiet and then got a secondary school recommendation after primary school. By luck and because my parents stood up for me, I ended up at a Realschule where I did very well in writing, but not quite as well in terms of speaking. After finishing secondary school, I wanted to become a nursery school teacher and had already completed work experience there. However, as I dared to be more open and talk more, I got a recommendation for grammar school. My high school biology teacher introduced me to biology, which made me engage more intensively with the subject.
Did you have female mentors or role models? And if so, how have they influenced you on your path (so far), what tips have they given you?
You could start with Marie Curie 😊. In my immediate environment, however, my female mentor was actually always my mum. It was always important to her that we, her children, chose a profession that would allow us to earn enough money to support ourselves and be independent of a man. She always believed in us. Later, my own ambition. And Susanne [Gebhard], she is a real role model.
Have you experienced any hurdles or bad experiences as a woman in the natural sciences? If so, how did you overcome them?
Not so much bad experiences. Hurdles more so, but regardless of gender. My teachers often pigeonholed me because I was very quiet, along the lines of: ‘She's no good at anything’. Sometimes also within the family. But sometimes men also labelled you as a bit of a dummy.
Assuming you had chosen a different subject/career direction, what would it have been?
Social/educational work, but also forensic anthropology. I once did a module in this field during my studies. It quickly became clear that I was interested in the subject, mainly to read about it, but not as a subject.
You have a lot of responsibility in the working group, even outside of research, what do you particularly like about your job?
There really is a lot. I really like the lab work. But I also enjoy training my students, helping others, organising, ordering things.
Would you like to remain active in the lab after your PhD or would you prefer a management role that reduces the work in the lab and shifts it to the office?
At the moment, although this may change, I can't imagine spending all my working hours in the office, I find lab work far too much fun for that. Even if it doesn't work out sometimes. I like the change between sitting and standing, between writing and experimenting. I think a 50:50 lab/office job would make sense for later. I can also imagine going abroad as a postdoc.
And last but not least, what tips/advice would you give to young female scientists?
First of all, don't let yourself be pigeonholed. Just keep going, even if everyone says no, you can't do it, just keep going, you can do it. Don't let anyone tell you that you're not worth it or that you can't do it. You can all do it. And trust your gut feeling, if it's right for you, then everything will fall into place.

BioTalks

Florian Hof

AlphaFold
We are still updating our wiki, stay tuned!

Florian Hof about Alphafold

What do you think about AlphaFold and how do you use it in your daily life as a scientist? Do you use AlphaFold often?
It is definitely a very good technology that fits perfectly into our time, with AI, and therefore has been leading from the beginning. I would like to emphasise that AlphaFold is a hypothesis engine, meaning that it is not the truth, which I think is a very good sentence to start with. I used AlphaFold a lot at the beginning of my PhD because it was quite new and I wanted to see how you work with it and what you can find out with it. And that's when I used it to do a kind of docking. So it more or less showed me two monomers and gave me suggestions on how they could bind to each other, which in that case was quite easy for AlphaFold because it's a conserved motif and a conserved pocket. There are several proteins that have already been crystallised in this configuration. Then I didn't use it at first because I solved the structure on my own. It was crazy how close the prediction was to the actual crystal structure. But it was also important to me that it really was a crystal structure, because the fine differences at the edge of the pocket make up the affinity differences and AlphaFold can't predict that, because it's still based on this homology principle in many parts.
Then I used AlphaFold again to think about fragments for a protein construct, how it is structured and how I can shorten some of it and what of it absolutely has to be in there because it is a structured area. The more structured the protein is, the easier it is to produce. I then also used AlphaFold to look up disordered regions that we can exclude, which are perhaps not so important for us.
Where do you see the opportunities or disadvantages of AlphaFold? You've already mentioned a few, but maybe you can think of more.
Well, as I said, I think it's a great hypothesis engine to get the first ideas for a structure or to get a feel for the protein you're working with. And also to use it to develop ideas about what you could research or how you can advance research into the whole thing, for example if you are investigating a bond between two proteins, e.g. with mutation analyses.
I find it problematic that people take it as absolute truth, that's a big mistake, especially among those who are not completely rooted in structural biology or have few points of contact with it, take AlphaFold as the truth and then present it as such. I find that very dangerous. That was also the first thing my doctoral supervisor told me, make sure it always says AlphaFold underneath when you show an AlphaFold prediction.
We've already talked a lot about AlphaFold. How does AlphaFold actually work?
The input is an amino acid sequence. Different sequences are required for dimers/heterodimers. AlphaFold then creates a structure on the basis of homology models and multiple sequence alignments, adds the physical conditions and then models the 3D protein model. Roughly explained; there are a lot of theories behind it and the neuronal network. This then runs in cycles again and again and is improved and refined, relaxed and restrained, and then you get the result with usually five models. This is also very important to mention, because it is not just because it is the first model that it is the right one. Model 0 (phyton programming, hence 0) is practically the one with the highest confidence, and then it is ranked. That's why it's important to look at other models and not just the first one, because just because it has the highest confidence doesn't mean it's the right one.
Constantin: I think AlphaFold has a significant advantage in that it can also look at the evolution of proteins via homology and not just one protein alone. This makes it easier to investigate questions such as the development of protein structures, for example through amino acid changes.
How do you usually access AlphaFold?
There are several ways. At the moment, AlphaFold 3 can only be used in our working group via the Google server; we also have an installation on the university's Mogon network, but you have to ask Google whether you can get the restrains of the model and that is usually too time-consuming. AlphaFold 2.3, which I used to calculate all of your predictions, runs via Mogon 2, the university computing server of the state of Rhineland-Palatinate.
What options are now available with AlphaFold? You said that you can predict monomers and multimers and do docking. I know that I once looked for DNA/RNA or small molecules.
This was not yet possible with AlphaFold 2.3. With AlphaFold 3 it is actually possible to model DNA/RNA or small molecules. Of course, this gives you a wider range of interactions, e.g. for better research into inhibitors. In any case, AlphaFold 3 can combine significantly more and much faster. It's easy if you compare the performance of a multi-billion company with our university computer. One minute and you have everything you need.
We have a spider silk protein that might aggregate itself. Do you think it's possible, if we give AlphaFold the amino acid sequence of a monomer several times as input, that you could also look at oligomerisation?
I think that AlphaFold can represent oligomerisation. Otherwise, of course, as soon as you talk about liquid condensates, it also goes in the direction of phase separation. Then you're also at the limit for solid aggregates, i.e. where you suddenly end up with aggregates if you're not careful. It's definitely worth a try. That's the advantage of AlphaFold, you can simply push something together and see what happens and then derive the prostheses from that, or try to.
Unfortunately, there are not many solved structures for spider silk. AlphaFold tries to work via homology. Do you think it is a problem to look at and analyse our spider silk structures?
Well, the training dataset is more or less the PDB (Protein Data Base), i.e. all the published structures. If there are not so many spider silk protein crystal/NMR or cryo-EM structures, then the confidence is probably just lower. However, I am sure that AlphaFold will still be able to do something useful or close to reality, but the confidence will probably be a bit lower. In other words, it cannot be ruled out that such structures cannot be analysed.
Disordered regions were also displayed in our protein. Do you think these were drawn in on purpose or simply because AlphaFold can't do anything with them (keyword: poly-proline stretches)?
I think they were added on purpose. The protein structure is then stabilised in any case, as the coil secondary structure is present in a string. I would also like to refer you to Katja Luck, who has done a lot of research on this in the benchmarking of AlphaFold.
How good do you think AlphaFold is when it comes to unstructured or disordered structures?
Not so good. It's simply the biggest problem we have with it. The approaches of Katja Luck, for example, more or less describe benchmarking and they have also carried out this fragmentation, i.e. they have made different fragments and then looked for domain-domain interactions, which are great to find, as well as pockets, but the motifs that sometimes lie in the IDRs are sometimes not so easy to catch.
What information can be derived from an AlphaFold prediction for the project?
The 3D protein structure, several models of it, different scores, PHI score, PLDDT score and you can use this to estimate how the individual residues relate to each other, and then you also get the confidence of the entire model.
And the scores then range from 0 to 100 %?
Exactly, from 0 to 100 %. They are actually percentage values. You can then transfer the scoring to the molecule so that you can see which parts are confident and which are not. The aeroplot, the green square, can compare residue by residue in terms of confidence.
Which analysis software do you actually use?
I actually use ChimeraX or PyMol to analyse the structure, although I prefer ChimeraX. One advantage of ChimeraX is that you can upload the aeroplot directly.
Do you have any additional tips if someone uses AlphaFold?
It's not the absolute truth. Just because AlphaFold says the protein looks like this doesn't mean it is. On the other hand, don't be disappointed if it doesn't look the way you think it will. You should always remember that AlphaFold is based on existing structures, which means it can't pull something blue out of the sky. In addition, AIs sometimes hallucinate. And, derive the hypotheses with carefulness, so be careful with what you hypothesise. Be aware that it's just a prediction. We can only really see the absolute truth if we use imaging techniques such as NMR, cryo-EM and crystallography.
How would you explain AlphaFold to your grandmother?
If you have a pasta pot and you put hard spaghetti in it and then let it cook, it becomes tangled: then you take the spaghetti out again with a fork. Then they have a certain 3D structure. There are programmes in biology that could practically tell you what the pasta shape looks like after cooking. AlphaFold doesn't do this with spaghetti, but with amino acid chains, which are our spaghetti, so to speak, and gives us 3D protein structures that end up with this shape.

Dipl.-Ing. Tim Gemünden

Topic title
Tim Gemünden is the managing partner of TT Holding, which is an organisation above his family business. Within the family business, they basically do everything that has to do with property, from land acquisition including planning and construction, with the special feature that they have their own staff for construction, i.e. they do a lot ourselves. The whole thing started around 140 years ago as a construction company. As soon as construction is complete, they also take over the letting, management etc. of the properties. Their focus is really on property management with the entire portfolio.

Prof. Dr. Susanne Gebhard

Title
Susanne Gebhard is a professor of molecular biotechnology at the Johannes Gutenberg University in Mainz and her research is on bacteria in general. They specifically have two research projects, one where they're trying to understand molecular mechanisms of antibiotic resistance and a very different second project where they are looking at biomineral formation by bacteria and how we can use that in sustainable construction materials.

Title of read more

For our purposes, it's best to stick with the topic of biocementation. So that's why I will ask you why it is so important to talk about the whole research of biocementation?
I think one of the big problems we have on the planet at the moment is our carbon dioxide emissions and you may or may not know that the cement production that we use to build our buildings is the second largest contributor to CO2 emissions. So, we really need other ways of making construction materials and specifically to replace cement. And interestingly, bacteria, almost all bacteria, are capable of making something that is very similar to concrete called calcium carbonate minerals. So that's things like limestone and bacteria can help make this. And so, if we understand how bacteria make calcium carbonate minerals. Then maybe we can use these bacteria to replace cement in at least some applications.
his process is called MICP. What is MICP exactly?
MICP stands for microbially induced calcium carbonate precipitation. And the calcium carbonate we mostly speak of as calcite and that's one of the crystal forms you can get. And this is basically a process that we don't really even understand all that well, where bacteria produce CO2 from their metabolism like we do. If there is enough calcium around, then somehow this process can lead to the formation of these calcium carbonate crystals. And we think that they grow on the surface of the bacterial cells, which act like little germ points where the crystal can start to form, but exactly how that works, we don't really know yet.
MICP is now used in many applications. We saw articles where researchers used organic fibres to improve the MICP process and the mechanical properties of the material. Do you know the current state of the science behind it?
I think most of the applications at the moment using MICP are things like consolidating soils. So, for example, river banks where you have a lot of water flowing past and wash the sand away. If you can get that hardened or the same with wind erosion or dust suppression and in general, where the MICP bacteria are put into the soil and then should make mineral there and then bind the sand to something a bit or the ground to something harder than it was before. And we ourselves have worked on a project where we put the bacteria in concrete. When we make the concrete, and then if the concrete cracks the bacteria can make some carbonate and seal the cracks.
One of the current really exciting fields of research is what's called Engineered Living Materials, which is the idea of using bacteria just completely as the cement. So you take sand, water, bacteria and whatever the bacteria need, and you mix it all together, and then the bacteria glue the sand together to make things like bricks. And at the moment, that kind of works. You get something of the sand sticking a bit together, but the strength of the material is really poor and it's currently nowhere near strong enough for application. There's some groups here and there who try to introduce things like polymers or fibres together with the bacteria to improve those properties.
MICP-based materials still have weaknesses. What are they?
I think, in the terms of material property the biggest problem is the strength. If you want to build a building out of bricks, you need to be able to stack a second brick on top of the first brick. Or if you want to pave a street or a road or something with stones, they need to last. And they don't currently last, if you make them just with the technology we've got available.
So we have decided to improve the MICP process by adding synthetic spiders silk proteins. Do you have any experience with spiders silk proteins or in general with spiders?
I don't like spiders. But I mean the spider silk is a fantastic material, but I've never actually worked with it by myself. My main experience is you cycle through the woods and the spiders webs gets stuck in your face and it's really hard to get rid of because it is very tough, wonderful material, but it is extremely hard to produce on an industrial scale. So I think there's a huge potential with spider silk, but there are also enormous challenges.
What was your first impression when we came up with the crazy idea to use spiders silk proteins?
Actually, it is a bit of a crazy idea to try to make this, because that's really difficult, as you know, but the idea is actually really good because we were looking for some fibrous material you could use in such Engineered Living Materials to help keep the sand and the mineral product together. Spider silk proteins got very strong tensile properties. So you can hang a lot of weight off a spider thread, but also it's sticky. And sticky seems like a good idea, if you want to glue things together. So I thought the idea is really, really good. But very difficult to achieve.
Do you think that with our idea we can overcome the weaknesses of MICP?
Yes. I think that you would have to make significant amount of the protein so that you can use it as a material. But that's not impossible. There are developments using spider silk in a bit more niche applications or smaller scale applications, such as wound dressings. Things like that, you don't need tonnes, but if you can produce a protein with bacteria, you can upscale that to industrial scale. It is doable in theory, you just have to find a good way to make the protein and in terms of properties for the cement, I mean we won't know until you've tried - But why not?
Speaking about production, we plan to use Bacillus subtilis as our host for the production of synthetic spider silk proteins, what do you think about it?
So I think Bacillus is often the good organism, if you want to make a protein, it's what Bacillus is good at, is what it's used for already. For example, if you think of enzymes for washing powder or other things that's it's done on an industrial scale. Bacillus is very amenable to industrial scale fermentation. There's a huge amount of experience, globally, using this organism. And if you want to get an enzyme produced and immediately exported from the bacteria, so you can harvest it from the culture supernatant, then Bacillus is really good because it's only got one membrane, so you only have one big permeability barrier.
Coming back to the spider silk proteins. As spider silk proteins are very large and repetitive in nature, we developed an own assembly strategy composed of Modular Cloning (MoClo) and RCF25. What advantages/strengths do you see in our assembly strategy?
I think the real challenge with your type of constructs is that you have a very repetitive protein, and I think in 2025 most people will immediately think of something like Gibson Assembly to stick big complicated bits of DNA together. And the problem is that Gibson Assembly really doesn't deal well with repetitive DNA because you end up with single stranded intermediates and misalignments, and then you will lose your repetitiveness of your DNA, which is exactly what you want to build up.
And the RFC25 is old, right? And it's often considered to be an out of date cloning standard, but it was specifically designed to stick protein coding sequences together in frame with just the minimal scar of two amino acids. And so I think using that for repetitive DNA because this is restriction enzyme based, it doesn't matter what the sequence is of the DNA. The modular design of the RFC25 means that in theory you can build up any number of repeats until your plasmid becomes unstable. I think that that's a real strength of the RFC25. I don't think that the little scar you always get between your repeats would matter because of the way the protein looks. And then you use the MoClo system to attach the elements that you always need to the outside, e.g. a promoter or a ribosomal binding site. That's really the strength of MoClo. And by combining the two assembly standards, I think you're really using them both to their best effect and you're getting rid of the problem of single stranded intermediates.
In the end of our project, we want to combine the synthetic spider silk proteins with MICP. Therefore we use two different bacteria, Bacillus subtilis and Solibacillus silvestris. Where do you see the biggest challenges in our project?
I think every time you try to combine two bacteria to do a job together, it becomes incredibly challenging because they are living entities and they live their own life and they don't particularly care very much what you want them to do. They are evolved in densely inhabited ecosystems, so they're very good at competing against other bacteria that want the same nutrients. And so at the moment, we don't know whether Bacillus subtilis and Bacillus silvestris would even grow at the same time, or whether one would simply outcompete the other and at the end of the day you just left with one of them. And so I think a really big obstacle is that you would have to get the two to grow together and harmonise. The other problem, if you want to go towards application is that your B. subtilis, the spider silk producer, is a GMO. You can't currently put a GMO in a product, at least not in Germany. You can in other countries, but it's always difficult. So I think a better strategy might be to use B. subtilis in the industrial fermentation where we know that it does really well to get the spider silk and then get rid of the DNA and the GMO from the spider silk. Then you have a product you're allowed to use and mix that with the Solibacillus. So I think that might be a more promising route to success.
At what point would you like us to be at the end of the iGEM competition?
I think, if you can actually make a synthetic spider silk protein in B. subtilis, I think you will have solved a massive issue. I think an easier task that I think is very achievable is that you can show that your assembly strategy works and that you can build up the complexity of your repeat number in your genetic construct at least. Ideally to see, if the Bacillus can make it. I think you should definitely get to the point of trying to get the facility to make it, but whether you have the time to optimise the process that we'll have to see and maybe you can try some simple experiments where you try to stick the spider silk and the biocement producer and some sand together and see if you can get the two together.
Will you or your working group take part in the competition after us again?
That's a difficult question to answer. I think iGEM is a fantastic opportunity for students and it's a lot of fun to work with all these enthusiastic students, who want to do this. But it's also a huge investment of time and money. So that would depend a little bit on how much buy in we can get from the university, from the faculty. But that of course also depends a little bit on how well you guys will do this year, because success breeds success. But I mean, you're the first iGEM team of this university and the university doesn't really know what to do with an iGEM team and all the bureaucracy is really complicated. And it’s very tedious to get money out of people. But, if they see what an iGEM team can do, and if you do well in the competition this year, then that might make things a lot easier. And in that case it might be very feasible to continue to support iGEM teams in Mainz.
Will our project be continued after us?
Well, the MICP side of the project, definitely, I think the idea of combining polymers with MICP to produce minerals is definitely an interest that fits really well to our existing research focus. Whether we do it with spider silk or not depends a little bit on how many problems you run into this here, if it looks really feasible - Then why not? It's exciting, even for other applications, it's a notoriously difficult protein. If you can make headway with that, that would certainly become an interesting project.
What do you personally like most about the iGEM competition generally?
I think it's the opportunities for students at quite an early stage in their training to really develop their own project and then see it through to completion and also with all the aspects around like learning how to design the project, but also learning how to ask for money, to communicate the science to other iGEM teams, but also the general public. And it's just an amazing opportunity for the students to really learn things that usually people don't even learn necessarily within a PhD.
And are there any aspects of item that you see more critically?
I think the cost of the competition is a huge barrier for a lot of students to participate in the iGEM competition. It really depends on whether you're at an institution that is able or happy to support the team. So I think students from less well of parts of the country or universities don't get the same chance as at universities with established teams and established funding for an iGEM. I think in in some cases, I think the idea iGEM puts a lot of emphasis on teaching entrepreneurship. But at the same time, it's very much about open science, and those two are very difficult to mix. I personally don't really see a problem with that. I think the research at this level should be open science and but it's still good to teach the students these ideas of entrepreneurship. But for some companies it might become difficult to engage with the iGEM competition and support the teams. If everything the team does is open science in the end. So that's a slightly tricky point to fix.