Integrating analog and digital modes of gene expression at Arabidopsis FLC

Article activity feed

1. 

   Version published to 10.7554/elife.79743 on eLife

   Jul 19, 2023
2. 

   eLife

   Mar 3, 2023

   Author Response

   Reviewer #2 (Public Review):

   1. Mechanistic details of how FCA regulates FLC have been extensively studied, and both transcriptional and co-transcriptional regulations occur. I understand that FCA affects the 3'end processing of antisense COOLAIR RNAs, which regulate FLC. FCA also physically interacts with COOLAIR RNAs and other proteins, including chromatin-modifying complexes, which establish epigenetic repression of FLC regardless of vernalisation. In addition, FCA appears to function to resolve R-loop at the 3' end FLC, and FLC preferentially interacts with m6A-modified COOLAIR by forming liquid condensates. FCA is also alternatively spliced in an autoregulatory manner, and fca-1 mutant was reported to be a null allele as fca-1 cannot produce the functional form of FCA transcripts (r-form).

   However, I could not …

   Author Response

   Reviewer #2 (Public Review):

   1. Mechanistic details of how FCA regulates FLC have been extensively studied, and both transcriptional and co-transcriptional regulations occur. I understand that FCA affects the 3'end processing of antisense COOLAIR RNAs, which regulate FLC. FCA also physically interacts with COOLAIR RNAs and other proteins, including chromatin-modifying complexes, which establish epigenetic repression of FLC regardless of vernalisation. In addition, FCA appears to function to resolve R-loop at the 3' end FLC, and FLC preferentially interacts with m6A-modified COOLAIR by forming liquid condensates. FCA is also alternatively spliced in an autoregulatory manner, and fca-1 mutant was reported to be a null allele as fca-1 cannot produce the functional form of FCA transcripts (r-form).

   However, I could not find any information on the fca-3 allele, which was reported to exhibit a weaker phenotype in terms of flowering time (Koornneef et al., 1991). In this manuscript, the authors showed that the level of FLC expression is lower than fca-1 and higher than Ler WT, but I could not find any other relevant information on the nature of the fca-3 allele. Given the known details on the function of FCA, the authors should explain how fca-3 shows an "intermediate" phenotype, which is highly relevant to the argument for an "analog" mode of regulation in fca-3. Therefore, the nature of the fca-3 mutant should be described in detail.

   We thank the reviewers for pointing out this omission. We have added much more information on the genotypes in the methods of the manuscript. We emphasise, however, that the rationale for selecting fca-3 as an intermediate mutant was empirical: namely, it generates an intermediate level of FLC expression (Fig. 1C and Fig. 1S1).

   1. The authors used a transgene (FLC-venus) in which an FLC fragment from ColFRI was used. Both fca-1 and fca-3 is Ler background where FLC sequence variations are known. I understand that the authors introgressed the transgenic in Ler background to avoid the transgene effect, but it is not known whether fca-1 or fca-3 mutations have the same impact on Col- FLC.

   We tested the expression of both endogenous (Ler) and FLC-Venus (Col-FLC) copies in these mutants by qPCR and found similar results (Fig. 1S1C,D), indicating that the fca-1 and fca-3 mutations have similar effects in both cases.

   1. Fig. 3A: I understand that Fig 3A is the qRT-PCR data using whole seedlings, and the gradual reduction of FLC from 7 DAG to 21 DAG was used to test the "analog" vs. "digital" mode of gene regulation in fca-1 and fca-3. I am not sure whether this is biologically relevant.

   Indeed, Ler is the only line that has transitioned to flowering during the experiment, with both fca lines being late flowering mutants. We totally agree that for Ler, later timepoints may be biologically irrelevant. It is used in this case as a negative control for the imaging, since FLC in Ler was already mostly OFF from the first timepoint and no biological conclusions are drawn from the later times. We have added a comment to this effect in the results section, also clarifying in the discussion that our focus is on the early regulation of FLC. Therefore, by looking at the young seedling in wildtype Ler, as we and others have previously, we are already looking too late to capture the switching of FLC to OFF. However, we expect that this combination of analog and digital regulation will be highly

   relevant to FLC regulation in wild-type plants in different accessions, partly leading to the differences in autumn FLC levels that were shown to be so important in the wild (Hepworth et al. 2020).

   3-a) The authors wrote that "This experiment revealed a decreasing trend in fca-3 and Ler (Fig. 3A)". But, I do also see a "decreasing trend" in fca-1 as well (although I understand that they may not be statistically significant). I also noticed that the level of FLC in fca-1 at 7 day has a greater variation. Is there any explanation?

   The level of FLC in fca-1 at 7 days is indeed more variable in these experiments. However, in a new second experiment, this is not the case (Fig. 3S2). In addition, a similar effect has not been observed in the ColFRI genotype (Fig. S9F of Antoniou-Kourounioti et al. 2018). Therefore, we believe this greater variation in one data set may simply be due to random fluctuations.

   For the decreasing trend in fca-1 in Fig. 3A, as the reviewer says, this is not significant. However, in the second experiment, we again see a decrease, which is now slow but significant. The decrease could be due to a subset of fca-1 ON cells switching off (in tissue that we have not imaged) and we comment on this slow decrease in the text.

   3-b) The decreasing trend observed in Ler (although the expression of FLC is already relatively low in Ler) may be the basis for the biological relevance. But Fig. 3D shows that the FLC-venus intensity in Ler root is not "decreasing". The authors interpreted that "root tip cells in Ler could switch off early, while ON cells still remain at the whole plant level that continue to switch off, thereby explaining the decrease in the qPCR experiment." Does this mean that the root tip system with FLC-venus cannot recapitulate other parts of plants (especially at the shoot tip where FLC function is more relevant)?

   The authors utilize the root system with transgenes in mutant backgrounds to observe and model the gene repression (transgene repression, to be exact). If the root tip cells behave differently from other parts of plants, how could the authors use data obtained from the root tip system?

   We now show that FLC-Venus in Ler, fca-3, fca-1 in young leaves have similar expression patterns to roots, thus validating the root system as an appropriate one to study the switching dynamics, see response to Essential comment 3. Nevertheless, in Fig. 3A, we show that FLC expression declines even in Ler. However, the levels here are low, so if it is indeed a subfraction of late-switching cells that are responsible, these cells cannot form a large proportion of the plant. We now make this clear in the text.

   1. I do see both fca-1 and fca-3 can express FCA at a comparable level (Fig. 3B); thus, I guess that the authors are measuring total FCA transcripts and that fca-3 may result in different levels of "functional form" of FCA. But this is not clearly discussed.

   We have now added yellow boxes in Fig. 2S3 to show additional examples of short files of ON cells in fca-3 and fca-4. To further improve the interpretation of this image (and all others in the manuscript) we have changed the presentation of the imaging using a different colourmap to enhance clarity.

   1. Quantification based on image intensity needs to be carefully controlled. Ideally, a threshold to call "ON" or "OFF" state should be based on the comparison to internal control and it is not clear to me how the authors determined which cells are ON or OFF based on image intensity (especially in fca-3).

   For the wild-type and fca-1 situations there is no switching in the model, and hence no dynamical changes in the FLC protein levels. As the FLC levels in the ON or OFF states are simply fit to the data using log-normal distributions, this would simply be a fitting exercise for fca-1 and Ler, and little would be learnt. Hence, we have not pursued this line of analysis.

   1. In many parts, I had to guess how the experiments were performed with what kind of tissues/samples. The methods section can benefit from a more thorough description.

   We have now gone through and added the missing information.

   Related to Public review #2. What is the phenotype (flowering time) of FLC-venus in fca-1 and fca-3? In addition, how many independent lines were used? Do they behave similarly?

   It was observed that with the additional FLC gene (in the form of the FLC-Venus), flowering is delayed as expected. However, this was not quantified in this work. Instead, we validated that the expression of the transgene was equivalent to endogeneous between genotypes, as shown in Fig. 1S1, supporting that this is an appropriate readout for FLC expression. One line for each genotype was selected and used in this work. In addition, we also now use fca-4, which has similar expression to fca-3, and where FLC-Venus also behaves similarly to the fca-3 case (Fig. 1S1, 2S3).

   Reviewer #3 (Public Review):

   1. The way the authors define ON and OFF cells sounds a bit arbitrary to me and, in my understanding, can affect a lot the outcomes and derived conclusions. The authors define ON cells to those cells having more than one transcript, or when they are above the value of 0.5 of the Venus intensity measure - what would it happen if the thresholds are slightly above these levels? And why such thresholds should be the same for the studied lines Ler, fca-3 and fca-1? By looking at the distributions of mRNAs and Venus intensities in Ler and fca-3 plants, one could argue that all cells are in an OFF, 'silent' state, and that what is measured is some 'leakage', noise or simply cell heterogeneity in the expression levels. If there is a digital regulation, I would expect to see this bimodality more clearly at some point, as it was captured in Berry et al (2015) - perhaps cells in fca-1 show at a certain level of bimodality? When seeing bimodality, one could separate ON and OFF states by unmixing gaussians, or something in these lines that makes the definition less arbitrary and more robust.

   As explained in Essential comment 5, we have removed arbitrary thresholding from the manuscript and only used absolute thresholds from smFISH (now changed to >3, and shown that our results are robust to varying these thresholds, Fig. 2S2). If all cells are in the OFF state and fca-3 just has higher noise/heterogeneity, then this does not explain the reduction in expression over time. Nor can such heterogeneity explain the short files of ON cells and longer files of OFF cells in Fig. 2S3: the cells should just be a random mix of varying FLC levels. Our results are much more compatible with switching into a heritable silenced state. Finally, with bimodality, this is difficult to see as clearly as before due to the wide levels of expression in fca-3, but we believe it is present: a well-defined OFF state together with a broad ON state. This broadness makes extracting the ON cells quite difficult as a completely rigorous unmixing of the two states is just not possible.

   1. The authors use means in all their plots for histograms and data, and perform tests that rely on these means. However, many of these plots are skewed right distributions, meaning that mean is not a good measure of center. I think using median would be more appropriate, and statistical tests should be rather done on medians instead of means. If tests using medians were performed, I believe that some of the pointed results will be less significant, and this will affect the conclusions of this work.

   Highly expressing FLC lines and mutants, such as ColFRI and fca-9, often used for vernalization studies, are late flowering, but do eventually flower even with no decrease in FLC levels (and so no switching). This is not an artifact of using roots versus shoots, and presumably arises from there being multiple inputs into the flowering decision which can allow the FLC-mediated flowering inhibition to eventually be overcome.

   1. Some data might require more repeats, together with its quantification. For instance, the expression levels for fca-1 in Fig 2E and Fig 3D at 7 days after sowing look qualitatively different to me - not just the mean looks different, but also the distribution; fca-1 in Fig 3D looks more monomodal, while in Fig 2E it looks it shows more a bimodal distribution. Having these two different behaviours in these two repeats indicates that, more ideally, three repeats might be needed, together with their quantification. Fig. 2C would also need some repeats. In Fig 1S1 C and D, it would be good to clarify in which cases there are 2 or more repeats -3 repeats might be needed for those cases in Fig 1S1 C-D that have large error bars.

   The data in Figs. 2C and 2E are both based on two independent experiments, with the results combined. The data in Fig. 3D is almost entirely based on three independent experiments. We have now stated this in the legend. The Venus imaging was performed on separate microscopes for Fig. 2 and Fig. 3 and this possibly accounts some of the observed differences. However, we do not think that the data in Fig 2E for fca-1 supports a bimodal distribution: the slight peak at higher levels is, we believe, much more likely to be a statistical fluctuation. For Fig. 1S1 C and D, we now clarify in the legend that n=2 biological replicates for fca-3 and n=3 for others.

   Also, when doing the time courses, I find it would be very beneficial to capture an earlier time point for all the lines, to see whether it is easier to capture the digital nature of the regulation. Note that the authors have already pointed that 7 days after sowing might be too late for Ler line to capture the switch.

   We agree that capturing earlier time points for Ler in particular is interesting and important. However, we have found that this requires specialist imaging in the embryo and we feel that this is really beyond the scope of this manuscript and will instead form the basis of a future publication.

   Read the original source
3. 

   eLife

   Mar 2, 2023

   eLife assessment

   Regulation of gene expression in many biological systems occurs either through a binary mode where gene expression is either on or off (digital regulation), or through an analog mode leading to a graded modulation of gene expression. In this manuscript, the authors report how these two regulatory modes are integrated into a one-way switch pattern to control the expression of the Arabidopsis floral repressor gene FLOWERING LOCUS C (FLC). They suggest that an analog regulation in the autonomous pathway precedes a digital regulation conferred by Polycomb silencing before cold exposure, and this temporal switch correlates with the strength of transcription at the FLC locus in different genetic backgrounds.

   Read the original source
4. 

   eLife

   Mar 2, 2023

   Reviewer #1 (Public Review):

   FLC is a gene involved in cold-dependent induction of flowering, as prolonged cold exposure leads to a progressive decrease in the level of this floral repressor as a result of a digital switch from an ON to an OFF state occurring asynchronously in cell populations. In this work, the authors analyze the contribution of analog and digital regulation to FLC expression in the absence of cold exposure. To do so, they use a genetic trick to be able to perform this analysis in the wild-type Ler ecotype where the molecular tools are available to do such an analysis. In Ler, an activator of FLC is missing due to a natural mutation and FLC expression is repressed during vegetative development by a pathway called the autonomous repressive pathway, allowing for a rapid transition to flowering. The authors used two …

   Reviewer #1 (Public Review):

   FLC is a gene involved in cold-dependent induction of flowering, as prolonged cold exposure leads to a progressive decrease in the level of this floral repressor as a result of a digital switch from an ON to an OFF state occurring asynchronously in cell populations. In this work, the authors analyze the contribution of analog and digital regulation to FLC expression in the absence of cold exposure. To do so, they use a genetic trick to be able to perform this analysis in the wild-type Ler ecotype where the molecular tools are available to do such an analysis. In Ler, an activator of FLC is missing due to a natural mutation and FLC expression is repressed during vegetative development by a pathway called the autonomous repressive pathway, allowing for a rapid transition to flowering. The authors used two mutant allele in one component of the autonomous pathway, the FCA gene. In the strongest allele, FLC is highly expressed and the plant are late flowering while in the weaker allele FLC shows a weaker expression and the plant requires an intermediate time between Ler and the strong fca allele to flower.

   The authors demonstrate that the expression levels of the FLC gene vary quantitatively in the three genetic background they use (Ler and two fca alleles), and that mutating FCA leads to an analog increase in FLC expression. The quantifications performed by the authors indicate that increased level of FLC correlate with a decrease in the proportion of cells that can switch OFF FLC, with the strong fca allele showing a negligible amount of cells that can switch OFF FLC. The authors further measure the half-life of FLC mRNA and FLC protein, and show that FLC expression switch from ON to OFF is a one-way-switching. They used these data to build a computational model of the regulation of FLC expression and show that the model can reproduce the dynamics of FLC protein level at the cell population level in a time-course with measurement at 7, 15 and 21 days after sowing. Taken together their work suggest that, at least in the weak fca mutant, a combination of analog and digital regulation of transcription explains the population-wide dynamics of FLC expression. The authors propose that this regulation could be explained by high level of transcription of FLC preventing the digital switch, as a result of the short half-lives of FLC mRNA and FLC protein.

   The finding of this work are potentially of wide interest to understanding transcriptional regulation by providing a functional link between the digital and analog mode of regulation of gene expression. However, the evidence of a link between expression levels resulting from analog regulation and the digital regulation are only partly supported by correlations from cell population-wide analysis of FLC expression. The authors did not provide experiments to more directly test that higher level of transcription could indeed prevent the OFF switch of FLC. It is likely but not shown that the ON to OFF switch of FLC is regulated similarly in the absence of cold exposure (this study) and upon cold exposure. Also, in their model, the authors used the assumption that FLC switches off at division but they do not test this important assumption. Finally it is unclear whether this combination of analog and digital regulation is relevant to FLC regulation in wild-type plants or is only relevant to the laboratory-induced mutants studied in this work.

   Read the original source
5. 

   eLife

   Mar 2, 2023

   Reviewer #2 (Public Review):

   In this manuscript, the authors proposed a mathematical model to describe analog and digital modes of gene regulation using FCA-mediated FLC regulation as a model. Previously, a similar approach revealed that the repression of FLC by vernalisation is digital. The authors utilized allelic variations of fca mutants (fca-1; a strong allele and fca-3; a weak allele), which resulted in the different levels of FLC de-repression. Unlike FLC in fca-1, where FLC is robustly ON or OFF states in the root cells, authors observed "intermediate" FLC-expressed cells (weak ON) in fca-3. The authors argued that these "intermediate" levels of FLC expression in root cells might indicate the presence of the analog mode of gene expression. In addition, the authors used the "age"-dependent FLC repression to validate whether …

   Reviewer #2 (Public Review):

   In this manuscript, the authors proposed a mathematical model to describe analog and digital modes of gene regulation using FCA-mediated FLC regulation as a model. Previously, a similar approach revealed that the repression of FLC by vernalisation is digital. The authors utilized allelic variations of fca mutants (fca-1; a strong allele and fca-3; a weak allele), which resulted in the different levels of FLC de-repression. Unlike FLC in fca-1, where FLC is robustly ON or OFF states in the root cells, authors observed "intermediate" FLC-expressed cells (weak ON) in fca-3. The authors argued that these "intermediate" levels of FLC expression in root cells might indicate the presence of the analog mode of gene expression. In addition, the authors used the "age"-dependent FLC repression to validate whether digital mode can occur in fca-3 and concluded that it does happen. However, digital OFF does not occur in fca-1, and the authors speculated that this might be due to a "high" level of FLC transcription. Based on these observations, the authors developed a simple mathematical model to predict the transition from analog to digital gene regulation at the cell population level. It is an intriguing model/conclusion to show the "analog" mode of gene regulation, and FLC regulation has been an excellent model system for understanding various modes of gene regulation.

   However, some significant issues need to be addressed.

   1. Mechanistic details of how FCA regulates FLC have been extensively studied, and both transcriptional and co-transcriptional regulations occur. I understand that FCA affects the 3'end processing of antisense COOLAIR RNAs, which regulate FLC. FCA also physically interacts with COOLAIR RNAs and other proteins, including chromatin-modifying complexes, which establish epigenetic repression of FLC regardless of vernalisation. In addition, FCA appears to function to resolve R-loop at the 3' end FLC, and FLC preferentially interacts with m6A-modified COOLAIR by forming liquid condensates. FCA is also alternatively spliced in an autoregulatory manner, and fca-1 mutant was reported to be a null allele as fca-1 cannot produce the functional form of FCA transcripts (r-form).

   However, I could not find any information on the fca-3 allele, which was reported to exhibit a weaker phenotype in terms of flowering time (Koornneef et al., 1991). In this manuscript, the authors showed that the level of FLC expression is lower than fca-1 and higher than Ler WT, but I could not find any other relevant information on the nature of the fca-3 allele. Given the known details on the function of FCA, the authors should explain how fca-3 shows an "intermediate" phenotype, which is highly relevant to the argument for an "analog" mode of regulation in fca-3. Therefore, the nature of the fca-3 mutant should be described in detail.

   2. The authors used a transgene (FLC-venus) in which an FLC fragment from ColFRI was used. Both fca-1 and fca-3 is Ler background where FLC sequence variations are known. I understand that the authors introgressed the transgenic in Ler background to avoid the transgene effect, but it is not known whether fca-1 or fca-3 mutations have the same impact on Col- FLC.

   3. Fig. 3A: I understand that Fig 3A is the qRT-PCR data using whole seedlings, and the gradual reduction of FLC from 7 DAG to 21 DAG was used to test the "analog" vs. "digital" mode of gene regulation in fca-1 and fca-3. I am not sure whether this is biologically relevant.

   3-a. The authors wrote that "This experiment revealed a decreasing trend in fca-3 and Ler (Fig. 3A)". But, I do also see a "decreasing trend" in fca-1 as well (although I understand that they may not be statistically significant). I also noticed that the level of FLC in fca-1 at 7 day has a greater variation. Is there any explanation?

   3-b. The decreasing trend observed in Ler (although the expression of FLC is already relatively low in Ler) may be the basis for the biological relevance. But Fig. 3D shows that the FLC-venus intensity in Ler root is not "decreasing".
   The authors interpreted that "root tip cells in Ler could switch off early, while ON cells still remain at the whole plant level that continue to switch off, thereby explaining the decrease in the qPCR experiment."
   Does this mean that the root tip system with FLC-venus cannot recapitulate other parts of plants (especially at the shoot tip where FLC function is more relevant)?

   The authors utilize the root system with transgenes in mutant backgrounds to observe and model the gene repression (transgene repression, to be exact). If the root tip cells behave differently from other parts of plants, how could the authors use data obtained from the root tip system?

   4. I do see both fca-1 and fca-3 can express FCA at a comparable level (Fig. 3B); thus, I guess that the authors are measuring total FCA transcripts and that fca-3 may result in different levels of "functional form" of FCA. But this is not clearly discussed.

   5. Quantification based on image intensity needs to be carefully controlled. Ideally, a threshold to call "ON" or "OFF" state should be based on the comparison to internal control and it is not clear to me how the authors determined which cells are ON or OFF based on image intensity (especially in fca-3).

   6. In many parts, I had to guess how the experiments were performed with what kind of tissues/samples. The methods section can benefit from a more thorough description.

   Read the original source
6. 

   eLife

   Mar 2, 2023

   Reviewer #3 (Public Review):

   Gene regulation at the single cell level can appear in two fundamentally different modes: a digital mode, in which a certain gene is either ON or OFF, and an analog mode, where a gene can gradually modulate its expression in a range of values. Yet, it is unclear how such two modes might operate together. In the work by Antoniou-Kourounioti et al, the authors argue that the Arabidopsis floral repressor FLOWERING LOCUS C (FLC) exhibits such two regulatory modes in the Arabidopsis root before cold exposure, with analog preceding digital.

   This work has the strength of performing an elegant combination of experimental and modelling approach to solve a non-trivial and fundamental question on gene regulation. At the experimental level, the authors are able to quantify the number of FLC transcripts as well as their …

   Reviewer #3 (Public Review):

   Gene regulation at the single cell level can appear in two fundamentally different modes: a digital mode, in which a certain gene is either ON or OFF, and an analog mode, where a gene can gradually modulate its expression in a range of values. Yet, it is unclear how such two modes might operate together. In the work by Antoniou-Kourounioti et al, the authors argue that the Arabidopsis floral repressor FLOWERING LOCUS C (FLC) exhibits such two regulatory modes in the Arabidopsis root before cold exposure, with analog preceding digital.

   This work has the strength of performing an elegant combination of experimental and modelling approach to solve a non-trivial and fundamental question on gene regulation. At the experimental level, the authors are able to quantify the number of FLC transcripts as well as their protein levels at the single cell level in the studied Arabidopsis lines, and they elegantly recapitulate some of their experimental results with an in silico root model.

   Although this work has a very high potential, I find there are several important aspects that require some attention.

   I think further explanations and clarity are needed to help the readership understand the differences between digital and analog regulation, beyond the explanations illustrated by Fig 1. In my understanding, digital regulation will involve observing some kind of bimodality when quantifying expression levels at the single cell level (see Bintu et al 2016), but from the definitions of ON and OFF cells the authors did in this work (see below), and the modelling they propose, it seems not to be the case. Given the authors derive very strong conclusions from their quantifications on what is digital and what is analog, I think it is important to be clearer in this regard. Also, to clarify the possible scenarios of interplay between analog and digital, I believe it would help to emphasize and better connect the modelling part to the experimental part.

   Another major concern to me is whether the extracted conclusions rely too much on certain choices the authors made when doing the quantifications from the experimental data. In particular,

   1. The way the authors define ON and OFF cells sounds a bit arbitrary to me and, in my understanding, can affect a lot the outcomes and derived conclusions. The authors define ON cells to those cells having more than one transcript, or when they are above the value of 0.5 of the Venus intensity measure - what would it happen if the thresholds are slightly above these levels? And why such thresholds should be the same for the studied lines Ler, fca-3 and fca-1? By looking at the distributions of mRNAs and Venus intensities in Ler and fca-3 plants, one could argue that all cells are in an OFF, 'silent' state, and that what is measured is some 'leakage', noise or simply cell heterogeneity in the expression levels. If there is a digital regulation, I would expect to see this bimodality more clearly at some point, as it was captured in Berry et al (2015) - perhaps cells in fca-1 show at a certain level of bimodality? When seeing bimodality, one could separate ON and OFF states by unmixing gaussians, or something in these lines that makes the definition less arbitrary and more robust.

   2. The authors use means in all their plots for histograms and data, and perform tests that rely on these means. However, many of these plots are skewed right distributions, meaning that mean is not a good measure of center. I think using median would be more appropriate, and statistical tests should be rather done on medians instead of means. If tests using medians were performed, I believe that some of the pointed results will be less significant, and this will affect the conclusions of this work.

   3. Some data might require more repeats, together with its quantification. For instance, the expression levels for fca-1 in Fig 2E and Fig 3D at 7 days after sowing look qualitatively different to me - not just the mean looks different, but also the distribution; fca-1 in Fig 3D looks more monomodal, while in Fig 2E it looks it shows more a bimodal distribution. Having these two different behaviours in these two repeats indicates that, more ideally, three repeats might be needed, together with their quantification. Fig. 2C would also need some repeats. In Fig 1S1 C and D, it would be good to clarify in which cases there are 2 or more repeats -3 repeats might be needed for those cases in Fig 1S1 C-D that have large error bars.

   Also, when doing the time courses, I find it would be very beneficial to capture an earlier time point for all the lines, to see whether it is easier to capture the digital nature of the regulation. Note that the authors have already pointed that 7 days after sowing might be too late for Ler line to capture the switch.

   If the above comments are addressed and the authors manage to clarify how the digital and analog regulation are integrated in the chosen system, I believe this work would have a strong impact on a very wide scientific community, given it tackles a very fundamental question in gene regulation.

   Read the original source
7. 

   Version published to 10.1101/2022.07.04.498694v1 on bioRxiv

   Jul 5, 2022