# **Genotype-Phenotype Meeting - Questions & Discussion notes**
[Website with program and meeting information](https://www.posnien-lab.net/dzg-gradmeeting/)
[Abstract book incl. participants list](https://www.posnien-lab.net/app/download/12054876131/Genotype_Phenotype_Program_Abstracts.pdf)
In this pad we want to collect notes and ideas coming up during our meeting. Everyone should have editing rights. Feel free to leave comments here (preferentially include your name)
## Questions for talks
We will copy all questions from the Zoom chat here.
@Speakers: Please check this space and answer the questions.
@all: Feel free to add additional questions here as you see fit.
<u>*Wed, Mar 2*</u>
13:00 Sonja Grath
- Natascha@Sonja: Is DNA methylation only not present in Drosophila and Tribolium or in all Diptera or Coleoptera?
- Nico@Sonja: IF DNA methylation is so diverse in insects, are there indications for diverse functions among those that have it?
- Sonja@Nico: Yes... I have to look for the original papers.. for sex-specific DNA methylation cool stuff has also been done by Laura Ross's group
- Gregor@Sonja: Methylation linked to genome size?
- Sonja@Gregor: Not that I know. There are insects with rather small genomes that also have DNMT enzymes. Hemimetabolous insects usually have more CpG Methylation than holometabolous insects, though.
- Natascha@Sonja: Are the d. Annanasse populations also different in other phenotypic aspects like pigmentation, size other morphologies? Do similar pathways/networks play pleiotropic roles influencing these morphological aspects and cold resistance?
- Sonja@Natascha: We started looking into phenotypes affected by oxidative stress and found differences that correlate with the cold tolerance phenotypes and there are shared pathways/genes for both oxidative and cold stress (also in comparison to D. melanogaster). We have not yet looked into other phenotypes (pigmentation, size, etc.) yet in D. ananassae.
- Luisa@Sonja: Interested in the CRISPR lines of D. annanasea. SUper cool to test candidate genes from Dmel in other species. So, you have couple (or one) line expresssing CRISPR-cas9? and, what are the complications you mentioned when actually editing the genome of that species? is it something particular about that species?
- Sonja@Luisa: We basically followed a protocol for germline transformation by the Gompel lab (http://gompel.org/wp-content/uploads/2021/04/Drosophila-transformation-with-chorion.pdf). Difficulties already started when the flies should lay eggs, Annabella came up with a specific plate based on grape juice that finally worked. Then, transformation was not really efficient. It is probably not only a problem of D. ananassae, but has to be figured out for each individual species. Take-home: Just that it has to be shown that Crispr/Cas works, does not make it necessarily straight forward :) Aother problem that has to do with the candidate genes might be that purely knocking them out already in the embryo, just might be lethal for the animal (we do not know the function in all developmental stages, for example) -> our next goal: tissue-/time-specific Crispr/Cas9 or genome editing in general
13:35 Ehsan Sanaei
- Luisa@Ehsan: what's special in your scale insects that they dont have the standard u-shape curve of infection prevalence? or is it a problem with estimation? it seems that your confidence interval at infection=1 is huge?
- Ehsan@Luisa: CI is my guess. The shape was supported by bootstrap values https://sfamjournals.onlinelibrary.wiley.com/doi/abs/10.1111/1462-2920.15833
- Nico@Ehasan: On which level the host shift defined? On species level? On Population level?
- Ehsan@Nico: Species level https://www.authorea.com/doi/full/10.22541/au.164544146.67062526
13:51 Alejandro Obispo Valencia
- Natascha@Alejandro: ISH in adults? Are you trying on specific tissues or sections?
- Alejandro@Natascha: We are doing in the whole embryo. We cut them in a half as we have premeabilization problems.
- Nico@Alejandro: Did you do whole body RNAseq? In which tissue(s) would you expect most obvious expression differences?
- Alejandro@Nico: We did RNAseq from central neuron system samples. Which is the tissue we suspect to have more differences due to the drastic behaviour changes.
- Natascha@Alejandro: genes with unknown functions? Conserved in other locusts? Gene duplications?
- Alejandro@Natascha: What I meant to unknown genes is that my group found assembled transcripts which didn't BLAST with any gene in any related (or not) species
- Nico@Alejandro: Just to get it right: The shift cannot be induced in adults, right?
- Alejandro@Nico: Yes it can be! After being exposed to gregarizing conditions they start to behave in a gregarious way (eat more, move more, meeting each other...) and this is maintained during the next generations and then is when you can see the morphometric differences -color and size-.
- Henrike@Alejandro: I missed, what processes are enriched in the differently expressed genes overlapping in both species?
- Alejandro@Henrike: We didn't reached that point yet! We are now in the assembling process, making the reference transcriptomes to move on in read counting.
- Luisa@Alejandro: do you think that mixing both species for the gregarious stimulation matters? what would happen if the stimulation is just within-species? could you explain the rationale behind that bit of the experimental design
- Alejandro@Luisa: I'm sorry I missexpressed that. We put both species in gregarizing conditions but not mixed together.
- Natascha@Alejandro: did you also tried to map against genomes of other locust species and check differentially expressed genes between species?
- Alejandro@natascha: As we are studying a typical gregarious species, and this species has not a propper genome assembling, we prefer to make reference transcriptomes rather to have the risk that the readings do not align and we lose information.
14:07 Somia Saadi
- Nico@Somia: What type of genes do you expect to be differentially expressed between solitary and gregarious?
Somia@Nico:type of genes which we expect to be differentially expressed between solitary and gregarious are genes of metabolisme , genes related with stress and meiosis .
- Natascha@Somia: Why are you using Abyss? Is it working better for this species than trinity? Or other assemblers?
Somia@Natacha: Im using AbySS, following the protocol of our group lab Bakkali M, Martín-Blázquez R (2018).
- Nico@Somia: Do you also have plans to check epigenetic changes?
Somia@Nico: yes, depnding to the results that we will have.
- Nico@Somia/Alejandro: Would you expect a few major genes or rather a change in expression of many genes?
- Sonja@Somia: The animal probably has a huge genome, or? I may have missed it..
Somia@Sonja No, this is why we will merge the transcriptomes of both phases of each gonad in order to obtain a reference transcriptome for each gonad.
14:23 Roland Zimm
- Nico@Roland: How do you distinguish between genetic and environmental causes for variation in tooth morphology?
-- Roland@Nico: There is no actual distinction between environmental and genetic factors per se; however, some parameters can be interpreted best as describing properties of the tissue environment (which, in turn, are impacted by some genes outside the core odontogenetic GRN kernel). Thus, we mainly group the parameters in terms of signalling factor parameters and biomechanical parameters, since the former correspond to purely genetic pathways, while the latter are more complex in nature.
- Natascha@Roland: Does the model allow to add/identify effects as canalisation?
-- I suppose canalization would not be something to be added, but probably an intrinsic property of the system. For once, directed selection for a specific phenotype successively reduces variation of mutants. In other terms, robustness increases or the populations get entrenched more centrally within the target region in the morphospace. The other thing is that phenotypes appear as "attractors", uniting surprisingly different parameters. I usually interpret this as very heterogenous factors interacting dynamically, with feedbacks, and indisentanglibly.
- Nico@Roland: What type of "variation" (e.g. TF binding affinity, protein function, cell-cell-interaction) is most efficient in expanding the morphospace?
-- I found parameters directly influencing epithelial growth and the negative feedback between inhibitors and activators as producing the richest variation.
- Luisa@Roland: when walking between different morphs (2 vs 3 cusps), does the model need large effect jumps in the parameters? or is a cumulative walk of small parameters steps?
-- In addition to exploration by random parameter mutations, I tried transiting between the two morphologies by gradual selection. I found, again, a difference in parameters associated with the change (yet, I need to quantify this statistically, the results are rather preliminary), meaning that it also matters HOW you walk inside the morphospace. Apart from that, the response curves differ between parameters, meaning that too small mutations may be quasi-neutral.
- YiranWang@Roland：why use sharks instead of mice? what mutation did you use? Thanks
-- Sharks have a higher number of teeth per jaw (and beyond, as their skin denticles represent a structure homologous to teeth) and show more gradual variation, so you get more "intermediates". Finally, we wanted to understand inhowfar chondrichthyian and mammalian heterodonty is achieved in different or conserved ways.
- Victoria@Roland: When a shark loses a tooth, can the replacement be a different morphology? And if so, how often does this happen?
-- Only on the long term (aging) or in the case of injury. The first teeth in hatchlings look conspicuously different from juvenile ones.
14:31 Emma Gairin
- Luisa@Emma: do your models recapitulate between-individual differences in pigmentation? meaning, is there space for variation in developmental noise, and genetic variation between individuals.
-- There are many different parameters that I can change (initial signals, rates of reaction, thresholds for reactions, strength of cell-to-cell interactions) to obtain differences in pigmentation and that yields patterns that could be found in individuals. My models are also randomised so never give the exact same results with the same starting patterns.
- Nico@Emma: Are the initial signals/coordinate systems different between zebrafish and Nemo?
-- In the zebrafish, the first interstripe is based on the position of the horizontal myoseptum (choker mutant with no myoseptum: no stripes). The interstripe appears more or less all at once along the body. In the clownfish the bars tend to appear on locations where the fish would have flexions for swimming notably (but there might not be any direct causal link) and they appear anterio-posteriorally (first bar to appear is the head bar, then body, then tail), so I would say that the initial signals are potentially quite different but they have not been identified yet.
- Roland@Emma: This is interesting. Would this imply that initial conditions might matter as much as gene interactions/diffusion rates participating dynamically in pattern formation? Which of the two allows for more evolvability/robustness?
-- The initial conditions matter to define the basic positions, and then the cell-to-cell interactions kick in to obtain the complex patterns. On top of that, growth might be important to control the final patterns in zebrafish, and that is something that I did not use in my models as we are lacking precise data on how the fish grow (their growth is not isotropic – they do not grow at the same rate in all directions. Overall the initial conditions, cell-cell interactions, and growth all play major roles in the pigmentation rates. An issue in initial condition leads to drastic changes in phenotypes (the choker mutant example) while changes in growth and cell-cell interactions would have more subtle effects and could probably allow for more evolvability and robustness.
- Sonja@Emma: Thanks for the great talk! very clear. Are there any other effects of pigmentation mutants on other phenotypes, for example, fertility or just sex differences in general? I remember that the clownfish is very special when it comes to sex determination and individuals can change sex.
- -- In clownfish, some pigmentation mutations are sometimes associated with other (unknown for now mostly) deleterious characteristics and many homozygous snowflake mutants for instance do not survive past a few days post hatching. In terms of fertility/sex determination, clownfish are initially male and can become female if the female of the colony die. These transitions occur at the adult stage at which point the pigmentation patterns have been established for a long time and are stable. Ecologically, having pigmentation / stripes / bars can help to camouflage / recognise conspecifics for instance so has impacts on fitness.
15:00 Luisa Pallares
- Gregor@Luisa: the allels found to be relevant only in high sugar - do they contain more CG genes than the loci at Standard conditions?
--> Great point Gregor, as I said I didn't test this specifically, but will do!
- Natascha@Luisa: Did you also test other environmental factors except diet? Or do you think that these would have similar effects?
--> Just tested diet because at this scale is not easy to make more than one experiment. But I do think it is important to test multiple environments to see whether there is common set of genes responding to stress overall, and which ones are actually specific to diet. In yeast, they have done similar things, exposing them to many different environments and found common set of 'stress-responding genes'. Will be great to see if this is also true in Drosophila.
- Sonja@Luisa: Actually a (female) fly DOES care about age and reproduction, right? Even if she is old, she still have offspring. So, accumulating mutations is not neutral - at least in comparison to humans where (at least female) reproduction usually stops earlier than the life
--> Agreed. This just points to the fact that theories of aging (mutation accumulatin and antagonistic pleiotropy) were clearly not developed thinking about Drosophila - all this theories come from mammals most of the time. Alhtough, if you take a look at Dmel fecundity curve, it peaks pretty early (first week) and declines super quickly after that. So the bulk of reproduction is biased towards early life.
--> Great, thank you. I also agree here. Further, the mating itself has an impact on the females (in Drosophila).
- Michal@Luisa: can you say more about the statistics involved in classifying SNPs into shared vs. cryptic? Depending on effect sizes and statistical power, some SNPs could be detected in one environment and not the other just by chance (without true GxE) and vice versa
--> we don't have independent T0 cages (ctrl vs hs) so we couldn't fit a full model with interaction effect (diet x time), so we run two models, one for each diet. And did a simple overlap to detect GxE vs shared. We did extensive simulations to see under which conditions (effect size, allele frequency, etc) could we be missclassifying lack of detection as GxE. The SI of the manuscript describes all this in detail if you are interested https://www.biorxiv.org/content/10.1101/2020.10.19.346312v3. In summary, given sample size in both treatments (thousands of flies and 3 replicate cages in each diet), and moderate allele frequencies (most of our SNPs), the number of SNPs that we could not be detecting in only one treatment is very small. Also, the effect sizes in both treatments are highly correlated which suggests that effect size isnot a big factor in not being able to detect SNPs in a particualr diet.
--> Thanks -- correlation between effect sizes, or some kind of variance partitioning actually seem like another way to globally quantify GxE. I'll have a look at the manuscript!
- Masse@Luisa: what was the method you used for sequencing high number of strains at the same time? was it a targeted sequencing or whole genome sequencing?
--> We used whole genome sequencing and the libraries were home-made based on Tn5 transposase tagmentation which makes library prep cheaper and faster.
- Katharina@Luisa: How do you physically track the timepoint of reproduction of such a large population to determine at which point a fly converts from fertile to the “shadow” until the end of life? And what is the reasoning behind stopping tracking at n=500 animals left? Wouldn’t including these give interesting insight into how allele frequencies change especially in animals with a higher longevity than compared to others?
--> We don't track fecundity, just time. So, for the modelling of the 'shadow' part of the lifetime, we just shifted the flat part of the curve over different time points and evaluated which one fitted better the experimental data. Regarding the 500 individuals left, they are actually included in the analysis, they are the last point that we included because as you say, they represent the animals with longer lifespan. We of course had to stop the experiment when enough flies were left (therefore ~500) to be able to reliably estimate allele frequencies.
- Roland@Luisa: Do genes whose activity depends on the environmental background rather increase or discrease variation in the phenotype?
--> We actually don't have a phenotype in this experiment!. We don't measure longevity per fly (because that will limit the number of flies that we could include in the analysis lowering the statistical power). So, we actually don't know how one particular gene affects longevity unless you test candidates one by one (we did that for couple of genes). BUT the question of increasing variance is a great one, and I'm particularly interested in that from the point of view of phenotypic robustness and evolvability. I'm quantifiying genetic effects on variance (not just on mean phenotype) using gene expression data on flies on different diets - manuscript coming soon!
- Nico@Luisa: Re: 31% of GxE loci. Do you expect similar trends for other phenotypes, such as morphology or physiology?
--> I don't know, that's why we need to test this at many levels of the Geno-Pheno map, including different phenotypes!. But my feeling is that the % might be lower in other more self-contained phenotypes, just because lifespan incorporates many other phenotypes so I guess the genomic space is larger.
15:35 Ekaterina Osipova
- Nico@Ekaterina: Any specific reason to search for loss of function mutations/gene loss rather than gene gain?
--> Several reasons here. 1) the concept of 'less is more' suggests that a gene loss can provide an adaptive advantage. 2) the previous screen for gene losses in fruit bats (also high-sugar diet) revealed several cool gene candidates. 3) We looked at a number of different genomic changes, including gene duplication (speaking about the gain), we just found a very cool candidate among gene losses and did the experimental validation.
- Nico@Ekaterina: Hypothesis why no stable CRISPR/Cas cell was possible?
--> Slow growth in general and sensitivity to the lack of other cells around. It is not specific to this cell line, it is also known to happen to iPS cells for example.
- Roland@Ekaterina: Does the loss of FBPases lead to an adaptive advantage by itself or does it rely, synergistically, on further mutations to be advantageous?
--> Definitely, there are other genomic changes contributing to the phenotype, including substitutions in some glycolytic genes. Based on the experiments, I would still say the loss of FBP2 alone provides an adaptive advantage.
- Victoria@Ekaterina: Do you know if anyone has sequenced other birds that primarily eat nectar, and if they have the FBP2 expression reduction? Or is it specific to the hovering adaptation?
--> Yes, there are other birds that feed on nectar, we are working on this project now :) Though I would say the loss of FBP2 in hummingbirds connects the evolution of nectarivory with the evolution of hovering flight, which is unique for hummingbirds.
- Roland@Ekaterina: I see that this adaptation is advantageous for the specific flight mode of hummingbirds, but an accelerated glucolysis would equally benefit small birds in general. So, is this an adaptation to small size first or to a lifestyle? And how about other tiny birds?
--> I would say, first of all, it is an adaptation for hovering since this type of flight requires huge energy flux and, probably, FBP2 loss could enable this flux from the continuously consumed sugars. Though, there is indeed a correlation between diet and the size of a bird: all nectar-taking birds are quite small, hummingbirds being the smallest. We are now looking into genomic changes in other nectarivors :)
15:51 Tânia Paulo
- Natascha@Tania: which kind of sensory organ development genes did you identify, can you by these already narrow down how flies could “sense” an infection?
- Luisa@Tania: I was thinking something similar. Can you elaborate how you ruled out that the adaptation also doesn't involve a) eating less food, b) eating less times per day. For example, did you test starvation resistance and related phenotypes?
--> Luisa, thank you so much for the question =) I did several experiments using CAFE assay and did not find differences in the amount of bacteria ingested. This was also confirmed when we did bacterial quantification at several timepoints throughout infection and found no differences in initial timepoints. They started to differ in bacterial loads later but already after immunity is activated so I would not attribute it to behaviour
- Maridel@Tania: you mentioned that gut permeability might be a factor in the adaptation: does that mean that there is variation in initial infection, or are all the hosts really getting equally infected and just differing in the ability to clear the infection?
--> hey Maridel! thank you for the question. We used multiple experiments (bacterial load counts and feeding quantification) to test for differences in initial infection loads and could not detect any between BactOral and Control. According to our data, adapted flies clear the infection faster and better.
16:07 Amanda Glaser-Schmitt
- Sonja@Amanda: Have you ever thought about comparing species or populations with different wandering behavior? I remember that some flies make quite some distance during this stage and wander a lot, while others are rather stational. Is this already set up during development and can it be seen at the transcriptional level?
-->Yes, this would be interesting. However, timing the species to ensure you are comparing the same stages would be difficult. Within and among D. melanogaster populations, there has been quite a bit of work on how the foraging gene drives differences in wandering behavior. In the populations I used, we also see a lot of variation in wandering behavior between and within the populations.
- Luisa@Amanda: why do you think late pupal stage is more divergent? does the expression pattern of different larval stages matches the 'developmental hourglass model' of less variation in certain developmental stages?
-->Great question, Luisa. I do think that it is related to the developmental hourglass model. I think that our finding that the late wandering stage shows so much expression variation indicates that it is not a developmental milestone and is less constrained than in the other 2 stages that we looked at.
- Roland@Amanda: Do you have any information about differential expression in the imago? Naively, I would expect population-level divergence to generally peak in later stages, since their effect on the adult phenotype should be more dramatic in early stages, unless metamorphosis acts as a variation buffer (in which case, however, comparison with other insects would be interesting)
--> Hi Roland, great question. For these populations, we only have expression data for these developmental stages. I think it would be really interesting to take a look at additional stages both before the examined stages (ie before the cessation of feeding behavior) and after the onset of metamorphosis to get an idea of how population-level divergence changes over developmental time. My guess is that population divergence would likely peak during the least developmentally constrained stages and be most constrained at developmental milestones. So basically in line with the hourglass theory of development.
- Natascha@Amanda: Did you rule out that the DEGs you find between stages dont represent heterochronic shifts of certain genes/pathways/processes?
--> Great question, Natascha. We controlled for this during the staging of our samples by using a combination of the blue gut method and the careful observation of known behavioral changes and stage-specific events for each stage in each individual isofemale line. So we are reasonably certain that our samples represent equivalent developmental timepoints for both populations. However, we did not check if there were potentially heterochronic shifts for subsets of genes or pathways during our analysis. I will look into this.
16:23 Saudat Alishayeva
- Natascha@Saudat: Do housekeeping genes actually really have a constitutive peak ATAC seq signal?
- Noa@Saudat: How do you know that your random reference sequence does not encode any regulatory information?
- Roland@Saudat: I see that you can already distnguish between meaningful and neutral mutations on the sequence level. Yet, there may be a lot of transcribed mutations that nevertheless have negligible effect on the phenotype. BTW, your last diagram is really reminiscent of the hourglass model.
- Nico@Saudat: Do you plan to confirm the positive selected sites? They may be related to gene expression differences?
- Natascha@Saudat: Might/could the gastrulation stages coincide with onset of expression vs. Maternal contribution
16:31 Arnaud Martin
- Nico@Arnaud: Do see indications that parallelism may be biased towards specific process (e.g. physiology, morphology...)
--> good q, we see a lot of parallelism in coloration and resistance traits even when we only keep the Linkage Mapping + GWAS studies (less bias).
- Natascha@Arnaud: Do you plan to also offer adding Gephe entries for experimentally identified mutations influencing phenotypes in the future?
--> see answer Luisa
- Natascha@Arnaud: Do funding agencies/Journals support the site, not only staff/funding wise but more by convincing grant applicants/authors to incorporate their results? Or is this automated?
- Luisa@Arnaud: good to see that lab-made large effect mutations are not there because they usually don't match effects from natural variation. But, do you plan on creating an interfase in the webpage to overlap this database with standard databases usually thought as documenting canonical genes based on lab-made modifications?
<u>*Thu, March 3*</u>
13:00 Hanh Vu
- Natascha@Hanh: which approach did you use to cluster gene expression, was it an unbiased approach clustering genes regarding similar expression patterns? Predetermined number of clusters or data dependent clusters?
--> Thanks for the question Nastacha! Before the clustering of genes based on their expression behaviors, my collaborators computed mutual information (MI) for each gene in our dataset. We then defined size-dependent gene as gene that has MI>1, which leave us with 300 genes. Our collaborator Steffen Werner developed a new clustering approach based on the network theoretical idea of random graph percolation to cluster the data. Please see “Werner, Steffen, W Mathijs Rozemuller, Annabel Ebbing, Anna Alemany, Joleen Traets, Jeroen S. van Zon, Alexander van Oudenaarden, Hendrik C. Korswagen, Greg J. Stephens, and Thomas S. Shimizu. “Functional Modules from Variable Genes: Leveraging Percolation to Analyze Noisy, High-Dimensional Data.” BioRxiv, January 1, 2020, 2020.06.10.143743.” for detail information. The number of clusters is not predetermined. It is defined based on the data.
- Juan@HanhVu: (I hope is not cheating if a lab member makes questions ;-), just joined). Question: Looking at the behaviours of the expression of genes variation with nutrional input/body size, I find the biphasic behaviour ones holding information about thresholds sensing (a switch). Therefore I wonder: have you tried to knock down any of these particular regulated group and see whether body size control is impaired? Eg.: well fed remain small
--> Thanks Juan! It is not cheating :) We haven't looked at the functional roles of genes that show "spike" behaviors.
- Nico@Hanh: Any idea about what genes/processes are correlated with different expression behaviors (up, down, spike etc.).
--> Thanks for the question Nico! There are still a lot for us to learn from the correlated gene expression behaviors and aspects of animal physiology. We do however, know few things about few groups of size-correlated genes. For instance, many of these size-correlated genes reflecting scaling of cell types with animal body size. For instance, gut markers increase their expression with body size while nervous system markers decrease their expression with body size. Some genes reflecting metabolic changes with body size (increasing in lipid storage in larger animals).
- Mette@HahnVu: What happens to the gene expression profile when naturally going from a large to a smaller worm after asexually reproduction. Sorry if I missed this in your talk.
--> Thank you so much for the question Mette. We only looked at gene expression profile after regeneration. We haven't looked at expression profile when the worm naturally decreasing in size after asexual reproduction. But if I could guess, I think a lot of the gene expression behaviors will be similar to regneration of the fragments that have similar size to fission fragments.
- Ting@Hanh: After the regeneration, is the size termination mechanism similar as the normal termination process?
--> Thank you so much for your quesiton, Ting! I think I might mis-understand your question in the first place. Do you mean how the regenerating fragments know which size of the tissue they need to regenerate and stop growth when it reaches correct size. This is a great question! Unfortunately, the mechanism that coordiate growth and size of tissue during regeneration still a question to be addressed in planarians.
- Roland@Hanh: Do, during regeneration, all organ systems increase size simultaneously or is there any distinguishable developmental sequence of events? By which mechanisms is the terminal body size defined, both for normal and regenerative growth?
--> Thanks for great questions Roland! We actually don't know if all organ systems increase or decrease (for instance the fragments with the brain, the brain will reduce in size to fit final organismal size) in size simutaniously or there is a sequence for the events during regeneration. We also don't know the mechanism that terminate growth in planarians. These will be problems for our group and other labs to address in the future.
13:35 Miriam Merenciano
- Sonja@Miriam: Have you also deleted the binding sites one by one or only all three together? (in the immune-stress reporter assay)
--> Hi Sonja! The experiment that I presented show the results of the deletion of all three binding sites related with immunity.
- Nico@Miriam: Are Dorsal, tin, DEAF-1 somehow involved in immune response?
--> Hi Nico! Yes, they are well-described transcription factors involved in immmune responses.
- Tânia@Miriam: Hi Miriam! Were the samples used for the qPCR of full flies or of just dissected epithelium? Also, did you measure survival of the flies that were missing the binding sites for the immune-associated TFs?
--> Hi Tânia! (Great talk yesterday, I really anjoyed it!). Samples used in the qPCR were dissected guts from non-exposed and infected flies. And no... we did not measure survival in flies without the binding sites. We just had CRISPR-mutant flies without the whole element. We deleted the transcription factor binding sites in transgenic flies, in which we microinjected a plasmid with the sequence of the transposable element in front of a reporter gene (with and without the binding sites).
- Tânia: thank you so much! I really enjoyed your talk too and you have super cool results =)
- Ekaterina@Miriam: How did you identify this particular (and very cool!)TE? Was this entirely from the literature or you ran some screens?
- Tânia@Miriam: one more question, how do you know that it is tolerance that is associated with the TE phenotype? Why not resistance? Thank you
--> Hi again! Maybe it was a bad choice of words. I said tolerant because flies actually got infected and died way faster than non-exposed flies. Do you think that it is better to use the word "resistant"? Thanks
- Tânia: Tolerance and resistance imply different immune defense mechanisms and they can both allow for higher survival upon infection. In broad terms, resistance implies that there is a reduction of the bacterial load by the host and tolerance means that the increase in fitness (in this case, higher survival) is independent of a decrease in pathogen load (probably there are metabollic compensatory mechanisms that allow for better coping). If you are not sure which of the processes is taking place I would go for a more neutral term like just saying "higher survival" or even "fitness".
- Miriam: many thanks Tânia for the clarification! From now on we will say "higher survival" as you suggest.
- Natascha@Miriam: Could the same upstream factors by involved in binding and driving expression of lime in adult gut stress and non stressed embryos?
--> Hi Natascha! It could be a possibility since Lime is a transcription factor, but unfortunately we have not checked that.
- Juan@Miriam: Is there informarion about the connection between Lime and Toll pathway? Have you synchronised the embryo collections for the embryo stage
--> So far, there is no connection between Lime and the Toll pathway. However, Lime gene was recently described as a link between systemic metabolism and immunity (Mihajlovic et al 2019). It seems that Lime is responsible for the mobilization of carbohydrates needed for the cell to get the necessary energy to face an infection. And yes, we synchronised the embryos for all the experiments. They were 4-8 hours-old in all the experiments shown.
13:51 Ting-Hsuan Lu
- Roland@Ting-Hsuan: Just to understand the developmental situation better: is eye size/ommatidia number a deterministic readout of size of the part in the respective imaginal disc that will give rise to the prospective ommatidia or is further growth (and/or later changes in the size of this area) involved as well? Is the relative size of this region different between the two species?--> Ok, I consider my question as essentially answered by later parts of your talk.
-->Hi Roland, what we know now is the ommatidia number is one of a deterministic readout of size. Although there might also other factors effect the final size during the pupa stage.
- Luisa@Ting-Hsuan: super cool to see the genes being added to the original network. Can you describe how that is done? based on GO terms? And, why do you think they were not found in initial screen?
-->Hi Luisa, the network is constructed by network tool "Genemania". They use a algorithm to systematcally add additional genes. What I understand is the gene are added according to the same datasets they used for constructing network (See answer below). In the initial screen, I focused on those differentially expressed between species or cis-regulatory divergence. Therefore, if they are not in at least two dataset I complied, they will be filtered out.
- Henrike@Ting-Hsuan: Thanks for the talk! I am also wondering where the data (physical interactions, expression etc.) for generation of the networks comes from? Which data base was mined?
-->Hi Henrike, the network tool I used is called "Genemania" and the dataset which used to generate the edges are from BioGRID, Gene Expression Omnibus and also some other protein interaction databases like PathwayCommons.
- Juan@Ting-Hsuan: Nice work! I wonder whether in your putative candidare analyses have you considered whether the change in eye size should be respect the proportional changes you observe betwen species (like the change in major and minor axes).
-->Hi Juan, that is a very interesting point. To get things clear, maybe it is better to discuss further in the coffee break. Thank you.
14:07 Linh Dang
- Luisa@Linh: wow! what is that gene in the volcano plot of DE genes that is super upregulated?
--> thank you for that good question. Because we are only interested in those DE genes having different open chromatin profile and SNPs associated to those genes, I have not checked that super upregulated gene. Yet I agree that it is interesting to have a look.
14:15 Gordon Wiegleb
- Natascha@Gordon: have you mapped the scdata to the species specific genomes or all to the same genome? Are they comparable enough?
-->I have mapped the data to the species specific genomes. This seemed to be necessary because the more complete annotation of the melanogaster genome was visible in the data when I used the same genome. They appear to be comparable enough now and I can find the same cell types in the datasets of all species.
14:23 Konstantina Filippopoulou
- Natascha@Konstantina: Do you think that the fact comparing holometabola to hemimetabola might influence the comparison.
-->That will actually be a great challenge,and this is why i intend to use more than one stage for these species
- Nico@Konstantina: How important will it be to sequence comparable stages?
-->It will be important to identify the stage where Nbs will proliferate with the higher rate, it will facilitate to capture the series of tTFs
- Luisa@Kostantina: how robust is the differentiation process. If you knockdown/knockout some of those TFs, is the process stopped? or is it able to continue with compensatory gene effects?
-->for drosophila we know that each factor activates the next,and inhibits the previous tTF (actually it is much more complicated finally), but it has been shown that the series cannot progress if you remove one. we actually want to check what that means for proliferation also
- Nico@Konstantina: Are there differences also in the final brain?
- Roland@Konstantina: How comparable are the involved GRN topologies? Is it conceivable that the particular roles of some TFs may have changed substantially (while not affecting the overall sequence of developmental events), which might bias the interpretation of observed differences?
-->This is true,and this is why I plan to analyze the genetic interactions among these factors, in species like tribolium and musca, where there are available molecular tools.
14:31 Sonja Prohaska
- Sonja@audience: any questions or comments? I am happy to answer them.
15:00 Nikola-Michael Prpic Schäper
- Nico@Niko: I think you mentioned it, but to be sure: Also the capacity to generate different phenotypes by reaction to environmental cues must be coded in the genome. Would you say that treating the "capacity" as phenotype would do the trick?
--> I´m really not sure... it still requires somehow knowledge about the "external trigger" to understand, which one of the possibilities will be put into practice.
- Roland@Niko: What strikes me about the examples in your talk is the apparent discrete nature of alternative phenotypes. Is this a example choice bias or do you think phenotypic "attractors" are more common than rather gradual variation?
-->The butterfly examples were probably "personal bias", but it is true that at least in the example of the Map Butterfly, the decision for one or the other phenotype is almost striktly binary. Maybe, if the "binary" phenotypes have their own selective advantages, this then plays the role of a "phenotypic attractor"?
- Katterinne@Niko: Hello, thanks for the very interesting talk! My question -> Do we have any understanding as of how an environmentally-induced epigenetic modification occurring during an organisms life-time is carried to its germline and therefore inherited to the next generation?
--> Not to my knowledge, but I have to admit that I am not an expert in epigenetics. But it is really a key question how epigenetic modifications of the "functional genome" are then also relayed to the germ line!
- Luisa@Niko: just a comment on the AI thing with millions of phenotype-genotype. That's being done in human and failing. We can discuss why..
- Nico@Luisa/Niko: Breeders do this as well with different success. In breeding population the predictive power may be better?
--> Yes, maybe because the "environment" during breeding is more controlled?
15:35 Maridel Fredericksen
- Nico@Maridel: Are the private genes "novel" genes or lost in one clone?
-->Maridel@Nico: We have evidence that the the overlapping genes (encoding putative FucT and WSC-domain-containing proteins) were lost from the resistant parent. Regarding the LTR retrotransposon, we did actually compare the F-locus region from several other D. magna clones from different populations, and none of them had this gene, so we think it was inserted recently in the lineage of the susceptible parent.
- Natascha@Maridel: it looks like whole parts are duplicated between the two, could the dosage of the genes also play a role? Is the coding sequence still the same or are the duplicates getting more different evolving to parlous over time?
-->Maridel@Natascha: Yes, there was a very large duplication in the resistant parent! But for both of the genes in that large duplicated region, one of the copies seems to now be a pseudogene (coding sequences are quite different, they seem to have a fragmented structure or fused with another gene, and no expression data). So, in the end we don’t have any genes that differ in number of functional copies between the two QTL parents, so it seems like gene dosage doesn’t play a role in this polymorphism.
- Tânia@Maridel: Thank you for such a great talk. I was wondering, do you know/have you tested if these genes would be important specifically against P. ramosa or to other pathogens as well? Also, what kind of functional testing can you do with Daphnia (sorry, I am not super familiar with the model)?
-->Maridel@Tânia: Thanks Tânia! We think these genes are specific to the attachment phenotype, which is important for resistance to P. ramosa but not other D. magna pathogens—the other pathogens appear to have different infection mechanisms, and resistance may depend more on immune responses once the pathogen has entered the host body, whereas in P. ramosa resistance depends largely on the initial infection step. In terms of functional testing, there was a lab in Japan that was able to use CRISPR/Cas9 in Daphnia magna. So far we haven’t been able to get it working in our lab, but hopefully soon (and/or with collaboration)!
15:51 Henrike Indrischek
- Natascha@Henrike: are there parallels to cave fish, where it was shown that eyes are lost or more stopped in development at different developmental “stages”
-->Partly. Cave fish blindness in is very recent, so much less mutations accumulated in comparisons to subterranean mammals. We looked at SERPINE3 in three blind fish species genomes and electric eels (+their closest outgroups), where there is a frame shift (so putative gene loss) in Lucifugu dentata.
- Natascha@Henrike: Mouse is also labelled in red, do they have such a bad vision or is this an artefact by the measure of sight?
--> They have low visual acuity, so this is not an artifact. There are obviously different ways to measure "vision" (eye size, color vision and also eye morphology might differ etc.), so with visual acuity, mouse does not perform well :) Mice can see, but obviously a lot less "well" than primates (who have fovea, color vision etc.).
- Nico@Henrike: The approach assumes that the number of loci that underlie eye/vision loss is highly limited. I assume that (developmentally) you can have multiple ways to loose vision. Do you see traits where this approach may be of limited value?
--> Also, I would say that eye is a very well suited system as due to the vertebrate WGD expanded a lot and there are many genes with a specific eye function.
- Natascha@Henrike: In the species with bad sight but without loss of Serpine3, is the protein seq different/ starting to become dysfunctional
--> Thanks for the question :) We did not check this, would be interesting to see for the blind fish. In general, SERPINE3 is rather lowly expressed in the species we checked und often is not included in the differential expression datasets, which complicates things.
- Natascha@Henrike: Is Serpine3 expressed in homologous eye structures in other animals?
-->In mouse and human, it seems be expressed in RPE cells, interestingly.
- Natascha@Henrike: Are the different candidates you found somehow connected? Could they belong to the same Network?
-->We did not build a network of the candidates, but I would expect them to cluster with other eye-related genes, possibly in rod- and cone-specific subnets.
- Sonja@Henrike: Perhaps also a bit related to 'loss' of eyes or vision. How would you actually study the difference between 'not developing an eye' and 'perfectly fine development of the eye, but later it is destroyed'. The 'phenotype' in the adult would be the same, but probably the reason and the pathways involved would be different?
--> Thanks for listening to the talk, Sonja :) We do not distinguish in our phenotype definition. But if the eye develops perfectly fine initially, I wouldn't expect many genes to be lost, so we would not detect gene losses in this species. Also, if the gene losses are very recent and the gene just accumualted a low number of mutations per coding gene length, we considered this as a "grey" gene in the screen, not as a "lost" gene. So it would not contribute to the association signal.
- Roland@Henrike: Building up on Nico's question: Might it be that in loss-of-function mutations, a list of genes becomes dispensible (as they were functional in, together, increasing acuity) and that even if the loss of acuity begins through a mutation in a different gene in every affected taxon, all affected species will eventually end up mutating/losing the entire set of genes? The rationale would be that the more complex or specific a trait is, the more loci (or developmental complexity or fine-tunning) would be required for it to be functional.
-->I totally agree. When more time passes and the genes involved in this specific process are not "used" anymore (=do not effect fitness), they are more likely to accumulate mutations and will eventually be lost. Still this happens by chance, so it is unlikley (or requires lots of time) to loose all eye-related genes in all blind animals. Another story are pleiotrophic genes (let's say a cilia gene involved in vision and reproduction), of course their loss would only occur if another gene could take over their role in reproduction and might involve remodeling of the gene network in expression and interaction. For our method, eye/vision is very suitable as we know that many specialized eye genes exist with probably low pleiotropy (see answer to Nico's question).
- Francesca@Henrike: Maybe I missed it or I am just not familiar enough with the technique, but how was the first genomic screen conducted? Were gene losses defined in comparison to closely related species?
--> Yes, they were investigated given human transcripts as the reference. Gene losses were called based on a whole genome alignment (WGA) of all the species to human for both, the initial screen and the more detailed analysis of SERPINE3 losses. WGA makes it possible to distinguish between gene losses and missing data and account for paralogs, which is essential for the analysis. If you are interested in the method, check out: https://tbg.senckenberg.de/hillerlab/tools-and-data/
16:07 Ellen McMullen
- Nico@Ellen: Maybe I missed it (soory if so): What is your explanation why the compensation kicks in in mutant, but not in RNAi experiments.
16:15 Memet Gözüböyük
<u>*Fri, March 4*</u>
13:00 Pavel Tomancak
- Roland@Pavel: You stated that animals are particularly excelling at "making shapes". What would you attribute this impression to? Human bias, richness of realized morphogenetic mechanisms (sensu Newman), their temporal, stage-wise, connection, absence of a rigid cell wall, early germline-soma separation, or evolutionary mechanisms (such as niche creation; although I would naively argue that immobile organisms such as plants should be even more forced to develop multiple strategies to deal with challenges in situ.
--> Maurijn van der Zee@Roland: the rigid cells of plants and fungi make morphogenetic movements much harder, of course
--> To be honest, I say it to motivate the rest of the talk which deals with animals. Also, the network of people we assembled around this ideas are all working on animals. I believe that in terms of shape diversity there would be other clades to consider. So yes, bias.
- Kristen@Pavel: What is the unit (scale of biological organization) that is subject to selective pressures, to specifically discuss morphogenetic modules as an evolvable unit?
--> I struggle with this question. I think I need to admit that I don't know. But I am convinced that we need to extend the definition beyond the information in the genome.
- Natascha@Pavel: Thank you so much, I completely agree with you that we need to study many more species! Do you have ideas/tips, how we could facilitate/improve studies on non-model systems given the lack of time/manpower/funding. So many young academics are advised to stick with the model organisms or just go that way because the work is less risky and thus more promising to get funded!
--> Well, my idea was to redirect some major scientific organisation into that direction to create a safe haven and source of resources for researchers interested in these questions and struggling in the current funding landscape. I am not giving up, although running out of options a little bit. Maybe something like Arcadia is the answer.
- Luisa@Pavel: wondering what are your thoughts on how important is studying ALSO genetic variation when studying morphogenesis.
- Maridel@Pavel: any idea if your suggestions would also apply to the plant world (i.e. is the time ripe to build such a bridge in field of plant development as well?)
--> For sure, also protists are an amazing playground to study morphogenesis on the level of single cells.
- Nico@Pavel: Gastrulation is diverse in insects, but the outcome is the same (extended germband). What would be the advantage of the different gastrulation differences?
--> This comes back to the developmental hourglass. The starting point of development and maternal investment is different in different species and gastrulation needs to react to that.
- Nico@Pavel: Re: morphogenetic modules. Do you see a link between morphogenetic modules and "gene (regulatory network) modules"?
--> I see morphogenetic modules as broader expanded definition of GRNs incorporating effector systems and physical interactions between cells and with environment.
13:35 Deepak Dharmadhikari
- Natascha@Deepak: do the isofemale lines overlap more with more closely related species or is it completely unrelated to the phylogeny?
-->Deepak@Natascha: Some isofemale lines vary more in one particular direction (intensity or size). But overall among 100 lines, there is a mixture of everything
- Amanda@Deepak: Sorry if I missed this in your talk, were all of your biarmipes lines collected from the same location/population?
--> Deepak@Amanda: yes. They come from one single orchad in Bangalore.
- Roland@Deepak: Sorry if I missed it, but what were actually the parameters or values the PC morphospace was based on? I also wonder how much inbred, lab populations without (explicitly) directional selection and natural populations would compare.
--> Deepak@Roland: Each pixel is an independent variable. i.e., approx. 270,000 pixels inside each wing. We look at overall variation in pixel values while preserving shape information. And with regard to inbred vs natural populations, inbred lines seem to vary less. Even *D. biarmipes* line that was used for Genome sequencing of that species varies less compared to isofemale lines, and they tend to occupy slightly different region in Morphospace.
13:51 Victoria Sharp
- maurijn@victoria: do the polyps also have eternal life when producing medusae?
--> No, they are only able to produce 2-3 medusae before dying.
- roland@victoria: Is is known what (or what kind of) signal from the symbionts triggers strobilation? What would the trigger be in taxa whose metamorphosis is not depending on symbionts?
--> It's likely to either be a trigger of the retinoic acid pathway in the host, or from signalling of protein ligands on the algae cells. In other scyphozoans, they do undergo strobilation, but the triggers are varying. In Aurelia aurita for instance, a change in temperature is the trigger.
- Natascha@Victoria: Do you plan to also restudy your approach in other more "stressful" conditions?
--> Not personally, but an undergraduate student of mine is!
- Nico@Victoria: Any idea about signals/molecules that trigger strobilation?
--> Based on the retinoic acid pathway, it seems beta-carotene produced by the algae is a likely trigger.
- Nico@Victoria: Any hypothesis why the different algae cultures had different effects?
--> My running theory is population divergence and the accumulation of somatic mutations across generations.
- Nico@Victoria: How does the algae profit?
--> They get protection from predation, and can use host waste products for a food source.
14:07 Maurijn van der Zee
- Natascha@Maurijn: Does Dorsal closure coincide with any large event on the genome level? I guess start of zygotic gene expression is much earlier, but any burst of chromatin accessibility?
-->I do not know, but yes, likely. At least it is known that ecdysone signalling induces major chromatin conformation changes.
- Nico@Maurijn: Is the pattern of associated SNPs maybe linked to structural differences (e.g. inversion)?
--> On chromosome 3, yes, i think so: there seems to be a big insertion even of several gene models between melted and the solute carrier. A pretty big structural difference.
- Natascha@Maurijn: Maybe I missed it, but are the non-selection lines also showing more differences starting with dorsal closure?
-->You mean if the starting population already had quite some variation in speed/start of dorsal closure? No, don't think so: the "normal distribution" seems pretty similar for the fast lines and the non-selected lines. Only the slow lines have a much broader "normal distribution" (possibly due to presence of a lot of alleles that all slow down development a little bit). So, it seems that with the selection of the fast lines, we really moved the population completely out of its original normal distribution.
- Natascha@Maurijn: Do you think that the TFs trk and br might have complicated combinatorial effects on regulation of the CYP18A1 or could the repeated seq. harbour some binding sites for proteins involved in chromatin modelling?
-->yes, the repeated seq contains one binding site for Tramtrack, which is an architectural protein, and known to be involved in ecdysone-induced chromatin remodelling.
- Nico@Maurijn: Is the "removal" of the repeated sequence "accurate"? May that be linked re-activation of TEs?
-->well, we sequenced the parents from every offspring we put into our Crispr-cas line in the end. It is not 100% clean, as we did not care about a few single bases introduced or removed extra, as long as the 230bp repeat was practically removed (so we rejected larger insertions, or other "damage"). So we kept a good control that the crispr-cas line is what we want.
- Matthias@Maurijn: Are you planning to insert multiple copies of the regulatory region to maybe get the opposite phenotype?
-->Nice idea, well at the moment I do not really intend to try inserting anything. Already pretty happy that we managed to get something out :).
14:23 Franziska Krämer
- Kristen@Franziska: nice, your DAPI for zen1 KO morphology is very similar to what I’ve had in the past by RNAi at those germband stages
--> Thanks! =) We are also very happy that the KO and RNAi phenotypes turned out to be so similar.
- Nico@Franziska: Are the effects on the embryo zen related or secondary effects due to loss of the serosa?
--> Yes, we also assume that the twisted germband and the altered position of the germband within the egg may be secondary effects. However, we will need to investigate this on a quantitative base. However, we think that the "big head" phenotype may not be a secondary effect.
- Nico@Fransziska: Do the ko embryos hatch? Comparison of RNAi and ko
--> We do not know yet whether the KO embryos hatch. The mVenus marker does not allow us to discriminate between hemi- and homozygous individuals, so we would need to do some genetic assays, which we have not done yet (we have the KO line for less than two month).
- Natascha@Franziska: Could your results of the twisted germ band/ bending posterior part after zen KO hint to a similar mechanism (crucial anchoring to the vitelin membrane) as Pavel just introduced
--> Yes, we assume that the anchor is missing in the KO embryos and this will lead to the twisted germband/falso position phenotype. We think that the effect is similar to the mechanism that Pavel has outlined earlier.
- Maurijn@Franziska: cool when you will have the homozygous line. May be a vulnerable line, though: probably more susceptible to desiccation or infections
--> Thanks! Actually, your studies have been part of our motivations to investigate Zerknüllt 1 via CRISPR/Cas9. If the line will be homozygous viable, we will consider checking desiccation resistance and probably also infection resistance.
- Georg@Franziska technical questions: I am impressed by your transgenesis efficiencies with HDR. Is this possible on a regular basis and how is this achieved? Maybe we can chat about some more details.
--> We have tested this "marker cassette insertion" approach only for this one gene yet, but for Zen1, we are quite content regarding the efficiancy. We invested a lot of time in optimizing the protocol, and it now includes a custom Cas9 source vector that extresses a codon-optimzed Cas9 under control of the endogenous alpha tubulin promoter. We would also be happy to share this vector (and our protocol and other resources) if you are interested!
## Discussion Day 1
*We would like to hear your thoughts about genotype-phenotype associations:
What is your model? What methods do you apply? On what molecular/phenotypic level do you study the G-P-map?
What type of phenotype do you study? Can you really distinguish a morphological phenotype from a developmental/life trait phenotype?*
--> Individual/Population (including sexual dimorphisms) phenotype --> Genotype
Both Genotype and Phenotype need to be properly defined (here are already different levels)
Definition phenotype: developmental biol. uses phenotype almost exclusively for mutant, but also behaviour, variation, epigenetic, gene expression
--> in the end when it vary it is a phenotype because it can be selected on
shift from gene-->trait to a more heulistic approach
bias of model organisms, we select for specific traits like being fast...
From Luisa Pallares to Everyone 05:31 PM
the article I mention is here: https://public.bibliothek.uni-halle.de/index.php/nal-live/article/view/1886
Arnaud: when people say “a phenotype” they always mean “a phenotypic state” and that is relative to a reference state, such as a WT state
Sonja: I would agree, Luisa, in the point, that 'having 4 legs' is not necessarily a phenotype. But still you can study how you can get 4 legs
Luisa: Right Sonja, that's what I meant at the end. studying how you 'make' something is also important. so, do we call 4 legs a phenotype? or, is it a phenotype just because it varies, some species have 4 and other 2 and other 100?
Natascha: exactly what varies both genetically and phenotypically totally depends on the angle you search, are individuals compared? Populations? Closely related species or distantly related species? But still are there general patterns during evolution that within a long distance protein changes/ gene duplications/losses are more likely than regulatory regions or is this process/gene/trait dependent or is it just a random mix of everything? Are there limits for certain traits or dev. Mechanisms—> like suggested by the ⌛
Other traits are more pleitropic/complex/connected —> I guess this is why studying whole networks and not only single genes is also important
Sonja: And you could come to the same 'phenotype' over different pathways (would have an example for that from lymphoma development)
Is not necessarily evolutionary biology anymore, but obviously still relevant
Roland: I agree, and lumping all genes or loci together as if they had a qualittively similar conribution to phenotypes is part of the issues that might make many studies complicated, Pleiotropical TFs and downstream effector genes are two very different bins
From Deepak D to Everyone 06:04 PM
To add to that, some Drosophila species look exactly similar but have undergone sexual reproductive isolation. So a different genotype produces similar morphological appearance
From Deepak D to Everyone 06:17 PM
Phenotype is a physical or biochemical response of inherited genetic material to a given environmental niche.
Phenotype is a developmental response to a given environmental niche.
From rolandzimm to Everyone 06:24 PM
no categories/no labels but we still need them and we need to connec them!
maybe we ought to confront students with problems, equipping them with nothing but heuristic tools and their intrinsic curiosity, rather than setting up disciplines
From Deepak D to Everyone 06:22 PM
Phenotype is a developmental
From Sonja Grath to Everyone 06:24 PM
My key point was: I actually do not need a label, but can be fine with NOT having to know everything
@PhenoDefinitions: how about taking advantage of the software-hardware analogy and shifting to traits whenever applicable?
From Natascha Zhang to Everyone 06:27 PM
I 100% agree we need to teach concepts and how to raise andnanswer questions and how to quickly find and evaluate literature
From rolandzimm to Everyone 06:29 PM
Plus, make the omnipresent JournalClubs interactive discussion clubs for anything except the topics that are directly relevant for your ongoing research
- snail, populations, physiological differences, salinity
- phenotyping + genome sequences
- reference genome available
- 6 populations , 3 +3, 6 individuals each for phenotyping
- phenotype & genotype can be measured differently
- out breeding us. non-outbreeding
- QTL mapping
- ultimate phenotype in fitness
- gene expression in more clear
- Mutation → phenotype Idev.
- us. population genetics point of view
- environment important
- "phenotype" needs clarification (mutant vs. population)
- stage- dependent effects important
- individual level is important
- stuck reductionist approach, traits cannot be studied independently leg. chemical)
- projects need to fit into a research career
- where do you fit in in an academic background
- phenotype-genotype micro-macro level
- QTL panels based on tiny snapshot, how relevant?
- models that are good to reveal candidate loci, large population panels to test for "relevance"
- specify phenotype → more & more detail
- specific may hide integration, eventually the function counts
- bulk RNA seq us. single-cell RNA-seq, too much detail is possible
- more generalist approach needed
- categories during studies already defined
- definition in important
- e. g. in light of epigenetics, defining the phenotype by inheritance comes too short
- a great TED talk that summarizes some of our discussion: [TED talk by Alejandro Sánchez Alvarado](https://www.youtube.com/watch?v=xsU10fX0aC0&t=758s&ab_channel=TED)
## Discussion Day 2
*We would like to open a few special-topic breakout rooms to bring similar ideas, methods and approaches together to discuss advantages, but also major limitations of a certain approach. Then we will try to bring the conclusions together in a condensed form to develop something like a roadmap for future collaborations.*
<u>What is your take home message from Day1?</u>
- Is there a uniform definition for Phenotype and Genotype and their relation?
--> already definitions and perception of key words like phenotype and genotype are very different between disciplines.
- A lot of research questions are driven by new methods/topics (faster, larger, more) instead of proper questions and appreciating and taking time to study descriptive details
- We know a lot, but biased (species, methods, lab conditions, mutants) far away from the reality --> holistic approaches are needed
- We need to define projects/research goals within broader concepts
<u>Discussion points/questions for sub-groups:</u>
**How can we fill the GAP?**
**How can we achieve a better understanding of Evolution/ Genotype --> Phenotype?**
1. How can we foster more holistic approches? Do we need to work more on general concepts than specific questions?
2. How can we further enhance scientific exchange across disciplines?
3. Do we need to change teaching/training the next academic generation?
4. How can we facilitate working with non-model species/natural variation instead of main lab reared model species?
5. more ideas?
- limitations: time and energy resources
- How do we start talking to each other? Maybe through the way one presents the own work. Place research in a broader context.
- current funding options often limit entering new fields
- Communicate big questions!
- generalist knowledge is needed to cross borders
- Importance of statistics/coding
- How to teach it?
- Learn with real own dataset
- programming language
- dataset + paper to show the end result
- Start with WHY?
- mathematical basis/theory of stats
- more real-life examples
- start broad, generalists first --> specialization happens within the project
- Question first. Choose the model based on the question.
- limitations in acceptance, funding, resources to establish tools/methods for novel systems
2. in person meetings needed,
- visit conferences/be part of graduate schools and try to connect to people you seem to like personally, collaborations across disciplines might evolve from there
- we need role models on all levels to ask questions out of your comfort zone, which might be considered as basic or "stupid", not be afraid to admit that you do not know sth.
--> students need to ask first, students organize seminars, special sessions with invited speakers and students (without PIs!)
- the barrier across carreer levels (PI, student, Phd student etc.) should be flat for discussuions etc., be more approachable
- still interdisciplinary discussion might work best in groups of one carreer level depending on your shyness/personality
- develop awareness of people from differnt discipline having a different "the language"/terminology and link, find same driver/goal
!!! We have the Solution: We need a time machine
1. Details still matter and are needed. All the Details are needed to find general concepts. But also integrate all information and find the rule, maybe AI coudl do the job. Collect and compare rules in details.
2. Maybe AI could be an unbiased unit finding connections, but AI again needs real data, so we need more studies from all ends.
5. Environment plays a bigger role of the evolutionary questions than that it is taken into account
Train Young academics? --> summer schools, teaching, international journal clubs?
Special issues in journals?
Collaborative topic confernces?
Think tanks of established researchers?
Special topic collaborative grants/Priority programs?
- Also interesting:
## Next steps:
- Meet @EED Naples
- Other DZG Meetings:
- Meet 114th @DZG Meeting in Boon mid September
- Join DZG Section Developmental Biology
--> Mailing list:
- Goevol Initiative: https://goevol.uni-goettingen.de/
- Priority Programme (SPP 2349) funded by German Science Foundation (DFG), starting 2022: "Genomic Basis of Evolutionary Innovations (GEvol)"
Interested in our work, please visit our Websites or follow us on twitter:
## Stand with Ukraine
###STAND WITH UKRAINE###
if you have a lab position/space or just want to help by sharing please announce it via this google Spreadsheet or the EMBO Solidary list
We would really appreciate your Feedback to the meeting, so please list all comments, ideas, future perspective and most importantly contructive criticism
Please send us an Email: