Capturing Emergence in Bacterial Biofilms | Research Lecture

[00:00:13]so yeah it's my my pleasure to introduce briefly prefix Sheraton who was speaker for today if you think it's a it's a special guy in many different ways and and I let you into that he didn't romantically busy over the past few years after he finished his master at the Institute of Technology in India he went to Singapore to NTU doing research there and then went to Amsterdam doing research there then he was here did two PhDs normal people just do one PhD but he couldn't stop so he did two PhDs one PhD was in anti-cancer drug efficacy and while he was doing that he was also setting up a company on that so he's a chief technology officer of a company or an anti-cancer drug efficacy models and then because he couldn't stop he also did research on biofilm and biofilm models and not only that also did experimental work both in the in the cancer research as well as in the bio film research and I guess he will talk about the experimental and the simulation work that he did in biofilm stuff so but a further ado the floor is your V fact we really look forward to your presentation thank you beta my screen so before I begin in case you want to ask any questions or point out something so if you click on your screen and you go to the top there will be something called view options you can click on the view options and click on an update so you can anonymously comment or underline any equations you want on the screen which I'm displaying you know so how representing about capturing imagens in bacterial biofilms so ice Peter said it's modeling an experimental work combined together so first of all a small introduction on what is biofilms so back here biofilms are formed by bacteria so we have motel bacteria which are moving around let's say in a liquid and all of a sudden they decide to settle on a surface after settling on a surface they lose their motility and they start to secrete something called EPS extra polysaccharides which are like a gooey substance which hold these bacteria together and give the biofilm a shape and after they grow to a some larger size what happens is they start to disperse which means they regain their motility go back to your new site and form and start a cycle of their life so if this cycle keeps repeating and the biofilms keep spreading on the surface so why would they lose their maitreya their motility what's the trade-off here so these biofilms they offer protection from antibiotics and fluid shear so basically if there is moving fluid it's better to stay within this biofilm community so they are shielded against any fluid shear also biofilms aid in nutrient accumulation so any glucose or nutrients surrounding this biofilm will be trapped inside this EPS so the bacteria can take the nutrients from the biofilm itself and finally one key key advantage of staying seda biofilm is cell signalling since they are nearby they can send quorum signals which with which they can communicate with other bacteria about the density of the biofilm or the presence of a predator or presence of a competing species within the same biofilm or biofilms nearby so why are they important why should we study them first of all they are the major reason for industrial corrosion so these are corrosion which we generally find in our household pipes and the second major reason is surgical infection so here is a needle and the biofilm growing over it so if you are going to use this needle on a patient it's going to result in a nasty infection and the most common thing which is dental plaque so the plaque which we have in our mouth in our teeth are basically bacterial biofilms and first of all the major reason why we got interested in bacterial biofilm is because the unique shape they assumed as they grow they form these mushroom shape structures which are flexible in the flowing liquid so they don't get detached easily so the natural question is what contributes to these mushroom shapes and if we are able to find why these mushroom shapes are formed will we be able to come with some chemical or physical method to remove these structures so that we can stop corrosion or remove dental plaque and surgical infections is it possible so we set out to seek an answer for this question so we published our work in AAC so we've worked with cell C complex T Institute in Singapore UVA itmo and USL so the first thing to consider in developing the battery or biofilm model is to model the individual bacteria so to model the individual bacteria we need to know the basic processes which the bacteria exhibit so first they move they lose motility only after they settle on the colony and even within the colony they are still able to make smaller movements inside the EPS the next thing is this bacteria how to grow and divide which is proliferation and finally these bacteria which keep on dividing how to adhere to each other so in our model and experimental studies we considered pseudo Pseudomonas aeruginosa bacteria and we modeled the bacterial biofilm to here to computationally model we implemented GGH based model cliche green aerobic base model so the reason why we chose this model was because it's a discrete model so here on the left you can see numbers 1 2 3 so basically these sites lattice sites indicate presence of a cell so one could be a single cell - could be another cell and three could be third bacterial cell so we have different energy minimization principles so in introducing these principles we will be able to get different shapes out of these virtual bacteria and this model is suitable for biological cells so as I previously discussed we need to model motility so we will be able to move these pixels around and we should be able to modify the growth so we will also be able to handle the growth and proliferation and finally we also have to capture the spatio-temporal effect so we shouldn't restrict ourselves just a GG H we also have to consider other factors such as nutrient diffusion and subsequent growth of the bacteria because of nutrient uptake so this model is suitable because this is basically computational grid on this grid they can implement any fvm finite volume method or finite element method base solvers to compute reaction diffusion equations so first of all I will talk about the mode lady model so let's assume the screen part here to be a cell so this is a boundary pixel of a cell and if a boundary pixel is to be copied here so we call it a copy attempt or in this case in biological case this would be called a membrane fluctuation so if we want to have a membrane fluctuations then our system should basically satisfy this equation so this equation indicates the change in energy sorry this is Del Delta e so the change in energy if it's negative if the new state formed is negative due to the state change from Sigma to Sigma i prime then your probability is 1 so the copy attempt from here to here will be successful however if the energy change is higher then it's dependent on this exponential here so here I have shown a trend of the exponential so it keeps decreasing with increasing energy so the the temperature term TM takes care of how much energy bandwidth or in struggle you can kill from within which a copy atom can be successfully made so random moment is possible only if this TM has a higher value and after defining the motility we move to the other two constraint which is the artesian and volume constraint so volume constraint is indicated by this equation Delta e V so in Delta e V if we are modeling the volume as similar to a mass spring model in which where we have the volume of the cell minus the target volume and lambda is more like a spring constant which keeps the size constant or makes the size approach the target volume so this target volume is determined from how much of nutrient is taken by the bacteria and what's the size of the bacteria computed by diffusion reaction equations which I will be talking in the neck few slides and here is an example of contact energy so contact energy basically gives a parameter which says if you have similar cells then what would be the hydrogen so the adhesion parameter is quite important because bacterias have different adhesion with each other and also to the substrate which is the bottom layer and if we combine delta e v delta ec then we have the probability change so this would be the change in total energy of the system and from this we will be able to know if a bacteria is going to grow move or adhesion adhere so this was the basic model which we set out with which we set out and thought like okay if we implement this model along with refuted reaction diffusion then we will be able to proceed further so now we move on to the biophysical model which I was talking about so these are the model equations so first of all we need to the bacteria need to have some nutrient to grow so in this case let's assume the major nutrient is glucose so oxygen can also be modeled as a glucose as a I salute so in this case I'm showing you an example of glucose so you have change in concentration of glucose is equal to whatever is diffusing minus whatever is convicted and finally whatever is react so in this here in this case the react term refers to whatever is taken up or uptake by the cells so this is the consumption term R so this R is defined by the mass of the bacteria B the metabolic coefficient which is the metabolism coefficient how much of the total concentration of glucose is used for metabolism plus whatever is used for the growth so mu M is a specific growth rate and why is the e in coefficient of glucose and K is the saturation coefficient here and then want to sell growth so cell growth this model by change in the biomass Delta B by delta T so here we have yield coefficient into whatever is total it's this R is the total consumption minus whatever is needed to keep the cell in balance or maintain the metabolic stasis of the cell so the model summary is we have cell motility proliferation adhesion convection if needed and nutrient uptake so so with this model we ran our preliminary simulations and this is what we observe so on the left you can see the simulation results so we sew them so this is a 3d simulation at the bottom we have files at the starts and the cells start to grow with time so 10 over 30 our 50 hour so as the cells grew we observed them more or less to be like a hemisphere rather than a mushroom shape however on the right you can see if we are going to take a wild type Pseudomonas biofilm then we have we observed a mushroom shape here so our model wasn't able to exactly replicate what was happening in the experiments so we we decided okay we are missing some crucial parameter which contributes for this mushroom shape structure development so there should be some or one or more than one term here which should be added to the energy equation to arrive at a mushroom shape so we thought about what would be the driving factor so we went back did experiments and we found that we had wild-type forming mushroom shapes and then we had BD la medial a is basically dispersion so if you remember my first life but I showed bacteria dispersing after they have formed into adult colony they started dispersed from one place to other so is that the driving factor for a mushroom shape to disperse so we rechecked it so delta b really so we made a dispersion mutant and found that they still phone mushroom shape so we decided okay dispersion is not a key factor then we went back and did a key wire mutant which is chemotaxis mutant so once we stopped chemotaxis we started to observe more or less hemispherical colonies so now we step back and we thought okay maybe we should put in chemotaxis to see how the model progresses so to model the chemos like chemotaxis we split it into three different cases so first of all when the saturation or chemo chemotaxis concentration is equal to zero then the net change in energy should tend to 0 which means the pixel copy atom should be successful and if the concentration change or if the change in concentration between the saturation coefficient and the current concentration is higher then the Delta e should be less than zero because we have the bacteria which are ready to move to a higher concentration region and finally if the concentration the saturation concentration is very high then your energy coefficient should be minimum and putting these three conditions together we arrived in a change in energy equation which is lambda which is the deciding parameter for how fast the chemotaxis is going to happen and concentration of oxygen here is old denotes the concentration of oxygen and this denotes the concentration of oxygen in I prime so here and here and now we have a complete equation Delta e is equal to Delta V C and one and this be maintained the same ggh equation so we ran the simulations and this is how the simulations proceeded so I am going to show you a video here so at the bottom you can see Phi initial see that cells and this is a 3d structure so this is a times time-lapse video as you can see the cells keep growing and the first structure you observe is more of a hemisphere but after some time you can see a mushroom cloud appearing so I will play it once more so dark blue cells are basically active cells and then you have light blue cells which are less active metabolically less active and you can see the final of mushroom shape structure so we went back and played around with different lambda chem value which is the rate or the how fast the cells chemo tax from one point to the other and we found that at different values we have different shapes of mushroom spawn so at a very high value we were able to find clearly differentiated mushroom structures so then we went back and we wanted to know what was driving this why is it forming a mushroom mushroom shape even those we introduced chemotaxis so we found that the motility at the bottom you have very slow moving cells and at the top you have incredibly fast moving cells compared to the bottom so the conclusion is that the top cells are the cells at the top the cells are moving fast and we found that the oxygen concentration along the cap was higher compared to the bottom so we went back and we plotted the results and we found that after time T is equal to 30 we have a rapid growth in the biofilm mask similarly we plotted the change in surface to volume ratio so we found at the start the the surface to volume ratio kept decreasing with time but again where we have the same inflection point at the same inflection point we start to see this surface-to-volume ratio keeps increasing which means the cells are not spreading so it looks like from 0 to 30 the cells were spreading on a surface and then all of a sudden they decided to move to the top and and after 40 you can see it's more or less saturating so we decided ok the first part from 0 to 30 is where you have the actual spreading on a surface then this point of inflection is where you start to form this next structure in the biofilm and finally from 40 you have the formation of the cap of the mushroom so this plot shows you the age of the biofilm in ours so you have bacteria very odd form that less than 35 hours at the bottom and after and the bacteria which have formed later are found at the top however what you can see here the most important observation as you can see small red cells may be consumed in young small red bacterial cells which are trying to climb to the top so what's happening is we saw a nutrient gradient here so until here we have lesser concentration of nutrient and here we have a higher concentration so these red cells which are motile and we want to survive they start to chemotaxis towards the top so that their population can grow and once they reach the top they settle and start to grow so this is the reason why we are seeing mushrooms shape structure so it's a competition based cell motility which results in formation of these structures and we see the similar case in experiments observe so until 35 hours we saw just a hemisphere similar to here and after that you can see multiple mushroom structures formed in the experiments so the model implications are we have three unique zones in the by bacterial biofilm form so the first zone would be the dormant bottom layer these zones are dormant because they don't have sufficient nutrients and at some point the cells here are going to die and get replaced by new cells and the nutrient limited layers of the stock so the stock is the neck region of the mushroom and here we have multiple cells forming which are in the nutrient limited zone and at the top we have the fast proliferating cell where the concentration is high concentration the nutrients is high so the cap keeps on growing so the problem with these kind of structures in antibiotic response is that the three different layers may respond in different ways to different antibiotics so there are antibiotics which target fast dividing cells and there are antibiotics which here which which target cells which are already and the hypoxia so a single antibiotic might not work on all these three regions so which could be a complication in a clinical setting so as such the entire biofilm has become highly heterogeneous over time so if we are going to remove the cap or the stock of the mushroom it will basically expose any dormant cells to the fresh nutrient supply so instead of basically removing or reallocating a biofilm structure we are actually helping them to revive to metabolically active state which is not decided in the clinical setting so in sitting in diseases such as cystic fibrosis where original severe films are formed the reversion of these dormant bacteria to active state could result to more acute infections so these were the model implications we observe just from analyzing how the structure is formed so what we did was now we have analyzed the structure so how would we be able to control or control formation of this structure or understand how any communication between this bacteria is occurring to to tell themselves about their neighbors their predators and they are and the flow around them so we did another study along with along with the same research group with some extra people so with cell C complexity Institute UVA it more UTS and which got recently published in environmental science and technology so in this setting what we did was we took individual so these are individual bacteria sorry these are individual clusters so we took bacteria as shown in panel C the green color cells so we Club them together in agarose and made them into smaller balls and place them as shown in panel D at equal distance from each other so this is our coordinate system so what it shows here is we have one secretor bacteria brand-new so in this granule all the bacteria are capable of separating quorum sensing signals so once the bacteria at zero comma zero secrete the signals they will start to diffuse outside and here at one point one point 5 comma 1 point 5 and 1 point 5 comma minus 1 point 5 we have something called quorum quenches so these granules secrete quorum quenching signals intracellularly and these intracellular signals basically quench any quorum sensing signals which are coming out all other granules shown here the blue granules are basically signal reporters so they'll light up once a signal reaches them so that's their function and we we used a camera to capture a time-lapse using a UV lamp and GFP filter and so this panel if shows what happens inside the granule so qsr signals are secreted and once they reach these cells then the cells start to light up and express GFP and if instead of cue signal Q s QQ enzymes come out then they will basically quench the Q signals so the cells won't be active and they won't express any G of B at the bottom you have we have the equation to model the diffusion of a HL so a HL here denotes the molecule which which is responsible for quorum sensing so the change in concentration of this molecule is equal to the diffusion happening plus any of the IHL molecule which is produced by the Q s granule - any of the molecule which are neutralized by QQ enzymes and anything - anything lost due to degradation and finally a bless term where convection happens so in this we have two different settings which I will talk about later so convection comes in a later path so I will play a video of how Q s in agarose happens so here you see a petri dish so at the center of the petri dish we have one secrete a granule which pumps out Q signals and around the dish we have the sensing granules which will light up once the signal reaches them so all our bacterial bio for granules so as you can see there is a circular diffusing pattern and all the granules are getting lighter as the diffusion progresses from the center to the outside so the file granule is here and it's lighting up and we have overexposure here so on the right we have panel showing different activities at different time points so 10 20 30 and 40 hours so on the right the red markings here in the experimental panel shows where we have secretor so the the video I showed we had just one secretor at the center so what we did was we found the activation times of these granules and we did our simulations and we arrived at the diffusion coefficient and we test we use this tuned diffusion coefficient to test different experimental setting and in one of the settings we have two different secretary annuals here so these granules after they separate you can see a pattern a cephalopod n of activation around them so the panel's B and C show you the activations in simulations and C shows the concentration which was observed concentration of the cue signal which was observed in the simulation and the activation pattern so panel D shows how close our experimental data agrees with the simulated data set results and panel a shows the mismatch in activation time which is how much of a difference is between is observed between experimental and simulator activation times so the maximum mismatch we found was around 3.5 hours so if there is a activation at 48 hours in the experiment then in simulation it could be either at forty four point five hours or at fifty one point five hours so that's what this mishmash time shows so our activation pretty close to what was happening in the experiments so now we have completed the Q s or just with currents in signaling so the next step would be to introduce a competing factor so the competing factor here would be Q Q which is the quorum quencher so this is similar to a biofilm setting where you have two different species and they want to send competing signals so that they can get nutrients before one another so to model this we used a similar equation to that of a HL except we have different production rates and also different degradation rate however we don't know the parameter values so the parameter values have to be extracted from the observation in the experiment very similar to what we did for Q s and this is the R aho just the rate of degradation of a HL q signal by Q Q and this is the rate of leak of molecules so before I talk about the leak so first of all I will show you the video so in this setting we have one sequitur at the center and two quenches at minus 1.5 and plus 1.5 on the right side so as you can see on the left you have all the granules activated however on the right you do not have multiple granules here and here activated that's because we have quorum quenches present here so these have started to separate any molecules which neutralize the effect so from this we have two crucial observations the first one is we have activations on the right side of the dish also which suggest that there are few q s molecules which are activating the granule even in the presence of QQ so the so the inference would be there is a delay in quorum quenching activity by the cube cube molecules so that was captured in our simulation and also this kind of this kind of quenching activity is only possible if any of the molecules QQ molecules leak out of the granule so that's why we added a leak term here so this QQ enzyme which is secreted within a material itself actually leaks out because of cell death and other factors and it starts to spread throughout the dish on the right side so you have quenching activity and now we go to the right side resource panel so at the top we have shown a different scenario so the red ring here so the red ring here shows the secretor the pink rings show the position of the quenching granule so after 20 hours we can see that the activation starts even the ones very near to the quencher are activated however the ones behind the quencher or not actually totally activated a few cells inside them on activated and and the bottom part is completely dark which means any progression won't result in activation of these granules here so we observe a very similar progress in our simulation also so we have we have activation green gfp activation here here and a very mild activation of the granules nearby and at the bottom we don't have any activations at the top and the bottom very similar towards occuring in experiments so here we see the concentration of Q so this part shows the concentration of Q s panel C so you can see that it's not actually spherical it starts out spherical spreading but as time progresses you can see the QQ activity overtaking qsr activity it's basically neutralizing so the concentration here is going down very similar to penalty where you can see actually the QQ enzymes getting out so ask QQ enzyme started spread then the concentration of Q s depletes so activation fails and at the bottom most well zfg PR we have show like different configurations in which our simulation was able to predict different activation patterns so here you have four different QQ granules together and here you have them in a line and then you have a pentagonal pattern and the large Pentagon to show different activation activation possibilities so the response is nonlinear in nature and with until now all the experiments were carried out in ocurro's phase so what the diffusion was moved and we had circular patterns but what the name in the next step what we did was we changed to an aqueous medium we basically shifted to water so first I will show the video of activation and with cube only with qs4 am signaling in aqueous medium so you can see that the activation actually starts from the edge of the dish and proceeds to the other edge of the dish and I will play the next video in queue as QQ and we can see that the activation again starts from one edge and progresses to the other Ridge and it's quite fast so in hewers we found that the activation in experiments took around 40 hours 48 hours our word in in aqueous medium we found that the entire dish was activated within 13 of us so this is not because of simple diffusion so there should be something more so we went back we we thought what would be the other parameters so we went and check the air flow velocity inside the sitting room and we came to a conclusion that the drop layer of the liquid in the dish was experiencing drag so whatever ambient flow was going around the petri dish was dragging the top layer and once you have a drag on the top layer then you have convection currents which keep carrying around Q s and QQ molecules and I want to show this Q s QQ once more to show that previously we had effects of QQ in agarose but in an equi setting even though there are QQ granules located at minus 1.5 plus 1.5 on the right side there is no quenching activity seen so Q s so QQ is not dominant in aqueous medium so our task here was to find the velocity and also it possibly explain what's happening in an aqueous medium setting so we modeled the this phenomena using conviction in addition to the diffusion terms we saw previously so we use the navier-stokes equation and we use no slip boundary conditions on the walls of the petri dish and a sliding wall boundary because you have are flowing on the top of the petri dish so your top layer is being tracked so that's the sliding wall boundary and then we have a slip boundary along the different layers and finally we have a pressure point constraint at 1 for one point of the petri dish it can be a random point on the dish so that it makes makes a tray moment in the petri dish so these are the results so first I will discuss about the Q s activation so in Q s we have in the experiment we have activation starting from the right side so what we did was our where B set of velocity vectors from left to right so as you can see the convection current takes whatever is secreted at the center it basically keeps moving it with time from three hour to six seven it basically took all the granule secreted from zero to three ever more it to the corner and then brought it back in so there is a current which is going from left to right and pushing the molecules the Q s molecules inside and once this molecule start to move they the current carries them to the other edge of the dish and they get activated the other granules so here you can see the actual concentration so whatever Q s is secreted you can see partial activations here the concentration is high in and at T is equal to 13 hours you have the entire dish which has become red important observation to look here is basically you have these granules which are white in color compared to the background which is blue in color so all these are 2d sections it's from the same level so why should the granules have more concentration compared to whatever is present on the liquid medium the reason is because the because of any of the accumulation which is happening inside the agarose bound granule so basically this is similar to the setting but I discussed EPS of the biofilm so this EP s is capable of tracking any nutrients from the surrounding so this agarose layer in the granule acts like EPS so it wraps whatever nutrient is present so in this case instead of nutrient it's actually trapping fewer signals that's why you have more concentration inside the cranium and it aids in faster activation of the granule so convection may be one factor for the faster activation but the other key factor for the faster activation in in aqueous medium is the retention of the signals within the granule so we went back and checked if the retention is true so we dropped one cue s q is granule in sorry one signal granule inside inside aho signals and after one hour we washed it we took the bait and we checked the supernatant and the beads and we found that the beads were able to retain part of the signal which we gave in we didn't know which signal was being given out so we did it with different molecules of a cho signal and we found that whatever be the molecule this beats were still able to retain the signal within them with time and we did the same for for for echoes medium with QQ quorum quenches and we found that the activation happened irrespective of the position or whatever the secretion amount of QQ r annulus so here you can see although the QQ concentration gets higher with time the Moloch the the granules there are already activated because the concentration of Q s has actually exceeded concentration of QQ and these ponds at an earlier time step [Music] so the conclusions we gathered from the study where for aqueous environments signals are rapidly disseminated because of the flow and they are received by distant responders so let's say there are two different competing species and species targets are very far away competitor then basically it can release its signal from let's say one meter away and it will be carried by convection to confuse the predator or another species at a faraway location and we found that QQ activity is not because of the intracellular presence of the q cube enzyme but it's because these enzymes actually leak out to the surrounding and that's what's contributes to the suppression of Q s effect and finally the convection facilitates ensure biofilm communication within the environment and the extracellular matrix actually retains the signal and therefore creates a competing Q s QQ molecules situation within the biofilm yeah so this is the thank you wonderful thank you so much Rebecca was again pretty impressive and and well delivered beautiful slide thank you I guess we have some questions we're all with a few of us so whom can I give the floor I'm pretty sure yeah pause sitting on the tip of his chair because this is very much in his interest okay let's click on that every Beck next talk there was something a bit in the second half these beats are they depict area covering them in a biofilm and is that why it's relevant that these beads capture the quorum sensing okay so we tried to create simulated environment very similar to what's happening so we captured so we grew the bacteria and kept them within an agarose gel so agarose gels tend to have a very similar property to EPS which is secreted by the bacteria and the biofilm so it's a pseudo boy of them which we form that's why we named them as granules instead of pseudo biofilms of stuff enough okay thanks beautiful talk very interesting so I was ruling about the confection which you use in your model I think discussed the direction of the confection that you investigated if this has an impact on your mushrooms or if it's influencing the growth form actually I did so I so since the tacos like 15 minutes so this was the best one so we had another study in which we actually analyze the flow and how the mushrooms were bending with with the flow and how they were sheared yeah we do have a study do you see any asymmetric grout forms due to flow for we did see so all so the there are David there are two parts one is basically a sheer study in which we did a fluid structure interaction problem now the one which is not published as basically how the growth was affected by this flow and if we found that still the bacteria tend to form these mushroom structures because at the bottom even with the presence of convection there was a concentration gradient between the bottom and the top so because of this gradient the bacteria was still trying to get to the top okay thank you thank you okay like one question I probably asked already um but I've got the answer so in the in the stalk growth that you showed us you see this this phase transition or this this point of inflection of 30 hours which is pretty steep and do we understand why it is so steep maybe you can show the slide again and so that we can think a little bit about it so so what's why is it so it really happens within ten hours or so yeah so at this point it could be so this steepness can be attributed to to two major things the first thing is at this point the concentration goes below our fixed set value here so in this situation you see SESAC which is the saturation concentration so it's a single concentration so at that point it triggers chemotaxis that contributes so that's why you have like one inflection point at which all the cells start to climb so if we are going to make this saturation parameter as a range instead of a single value maybe we will see a larger spread where we have this instead of 10 nos we would be saying 15 hours and another key point is at this time we see the formation of the of this neck here so this talk so this talk basically goes in our vertical or in a 90-degree fashion so the only way the cells can rearrange that fast in the model is to stack them one over the other which means you have more surface area increasing rather than occupying more area so that also contributes to this steep increase in time increase in the change so so one part of that research was also trying to investigate how to get rid of biofilm right yeah whereas of course Transkei and and Yap are very much interested in keeping them but and and then you said ok you know because of the different metabolisms you need different ways to get rid of of this of these mushroom shaped organisms but what about other things like you know like I know you can denature the whole thing with with soap or but it may be high temperature anything known of the differential effects of that like so if I would increase the temperature would then you know the top of the stalk first remove be removed or I don't know so how do you have any idea on that ok thinking about temperature it's again basically going to set gradients within the biofilm so the bottom part is going to be more shielded from the temperature effects than the Tom or the surface which is exposed and the problem is with this removal method with these removal methods is the practicality so we can't apply such methods inside a patient's body we can't pump them with bleach and also in pipes corrosion like in boilers in industries we can't just use some chisel or tools to just basically scrap them off so that's one okay okay clear yeah so we need to tell Trump about this so any other question any hands raising okay if not then thank you so much again the feedback for a lovely talk and as this has been recorded we can you know invite our colleagues to have a look at it because I think it was definitely much it definitely worth our while so we will put it online and share it with other people thank you all for joining us and I hope to see you I understand from done way that we will meet again after the summer break and when will there be approximately yeah so we will start again around September's okay so stay happy stay healthy I hope