 So we'll come again for the afternoon session. As I told you, we have a bit to reschedule this session. I will give a few slides to address the problem we have actually discussed on Monday and also in some of the discussions on about biology, reproducibility, and biology and a few thoughts about assumptions we make just to stimulate some discussions. So this is from our work. I've mainly worked with mammalian cells. And I just want to say very often if we do flux analysis, we do steady state assumptions, right? All of us do. We have people working with prokaryotes, but also people working with all karyotes are doing this. So you look, consider the growth with a lag face. Then you consider the exponential phase. And there is a transition phase and a stationary growth phase. And usually for metabolic flux analysis, this stage here or this phase is considered stationary. OK, this is for a terren cells, which takes some time to attach to the surface and then grow. If you're lucky, you're working in mammalian cells with suspension cells. Here you have the option to go in through chemostat culture so you can achieve a true steady state and do calculations. Then I would like to talk a bit on batch to batch variation because this was also discussed on Monday. And let me start with these quasi steady state assumptions. MDCK cells is a doc cells, doc kidney cells we use for influenza vaccine production. If you look at growth pattern, you have an increase in the cell numbers here. And then there's a stationary growth as they are confluent on the surface, in a microcarrier or in a T-flask. So if it comes to this phase here, this exponential phase, in these cells there is an early exponential phase with true exponential growth. But then soon the surface gets limited and they cannot grow exponentially anymore. There's a transition here, rather long tradition phase and an intermediate phase and then the stationary phase. And a lot of people, including myself, we have to define quasi steady states if we want to do flux analysis. And then usually we go into this phase here. You see here this is the glucose concentration. This is the glutamine ammonium and so on. However, this looks nice and in a lot of publications that are considered only the exosolium metabolites. But if you look into the cell, the situation not as simple here. In this phase, there's a steep decrease in the glucose 6-phosphate concentration. Look at the succinate, for instance. There's an increase in the steep decrease. This is dropping. This is increasing and dropping. Again here the complex dynamic on malate. So the steady state you can define based on extracellular data. And usually you take a rather large interval here. You have just to be aware in the cell the situation can be really far from stationary. If you make an assumption that you're risking that if you take the wrong time window, you're facing non-steady states in the cell. OK. And then towards suspension cells, you can do continuous cultivations. You grow the cells as exponential growth. You switch here to the supply of medium to continuous cultivation conditions. They readjust. And here we did a glucose pass. Don't worry about this. So you have an exponential phase and you have a steady state phase here. And same here for the glucose and for the lactate concentration. Here, first of all, the glucose is taken up, the limiting substrate, steady state. Lactate accumulates and is washed out and also here it's steady state. And now you can say that's ideal. So I take all information here from the steady state for metabolic flux analysis. However, if you do this, you have to be aware. The results you get here for the continuous cultivation versus the exponential growth phase is not the same. In the continuous cultivation, the cells switch from high-close glucose to low-glucose metabolism. So it's a switch in metabolism. Reduction glutamine uptake, ammonia release. I can show you the drastic changes in amino acid metabolism. So even if you achieve steady state conditions in continuous cultivations and you do flux analysis, be aware that an experiment in batch is not necessarily giving the same results as an experiment in continuous cultivation. There are significant differences here. And then we have the initial problem. If you want to reproduce experiments in mammalian cells, for instance, this is an experiment. This is a second experiment. Dilution rates almost the same. And if you, for instance, now analyze lactate uptake or glutamine glucose uptake, you see what the cells achieve here is multiple steady states. So it's very easy if you do mammalian cell cultivation that even if you try to achieve a very, very high reproducibility, every single experiment you do, you achieve multiple steady states. So two different groups is a good chance that they get different results. They can discuss forever. This is known from our group. It was done also in other groups, also by Stephanopoulos, for instance. So in biology, there's a good chance you get multiple steady states and these multiple steady states don't make analysis easier. And then we have this problem in reproducibility. This is where experiments where we wanted to see is there a difference whether we cultivate cells in a still tank biorector or in a wave biorector. We did quite a few cultivations in still tank and also in wave biorectors. This is all the accumulated data here. And you see there's some variation. Actually, there's quite a lot of variation. And we analyze the data and we described a metabolic flux analysis approach to differentiate between these cultivations. And then the reviewers were not happy. They say, okay, you have no clue how to handle my male gen cells and your experiments are just crap. And then they say, okay, what can I do? But then luckily, Genentech published this data here. This is data from male gen cells growing under, I would say, the best standardized conditions we have under GMP conditions. And here you see the same story. Here is a large variation between the batches. So this is supposed all the same, yeah? The red ones are ones where you have a different steady state. So these cells, also they are supposed to behave as these cells, have a different metabolism. And you see the large variation if you analyze biology. This indicates also the range you have to cope with. And then if you look for batch to batch variations and we consider the assay errors, errors we do in calculating this metabolic flux rates, biological replication, and so on. The overall errors here on the rates in our experiments go from 7.54 here to glutamate, 80.2%. Yeah, so there's a significant error you do in calculating these flux rates. And if you now compare these batches we've done in a similar growth rate here, specific growth rate range, you see there are significant variations here. The glucose in this experiment, the rate is 203, this 290 in the wave. But if you look for lactate for instance here, it's 300 versus 600. This is supposed to be the same experiment in the same biorector. And the same holds for glutamine is a factor of two between biological replications. So you have to be extremely careful if you interpret metabolic flux analysis results. So my summary here would be this inter-experimental variance for experiments performed under seemingly identical conditions. However, with different pre-cultures is a major contributor to variance. The differences we see between these biorector systems are statistically not significant and insufficient to confirm differences between cultivation systems, even if they look large. And finally, if you have a limited set of experiments, how many experiments would you need to confirm a difference between both systems? And if you do some statistics, we can say for the data I've shown you, we would need at least seven biologically independent experiments to separate the performance of the saline in a stir tank versus a biorector. And tell me who is willing to do these number of experiments, probably nobody. The maximum number I see is three, if I see a three at all. So it's just a warning. Biology has a lot of variation. It has multiple steady states. And unless you do a lot of experiments, probably you will not be able to see or to validate differences between cultivations. Okay. Now I switch to the first speaker, which is Erika Tarcano. She will talk on synthetic biology for the production of high-value chemicals.