Variable paralog expression underlies phenotype variation
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
In this elegant genetic study, Bailon-Zambrano and colleagues draw on classical genetic concepts to address the clinically pertinent question of how genetic variants in the same gene can yield wildly different phenotypes in different individuals. From their case study they conclude that a major contributor is variation in paralog expression. The question addressed is of great interest to evolutionary and developmental biologists in general and to those studying the evolution of developmental mechanisms in particular.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript.The reviewers remained anonymous to the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
- Genetics and Genomics (eLife)
Abstract
Human faces are variable; we look different from one another. Craniofacial disorders further increase facial variation. To understand craniofacial variation and how it can be buffered, we analyzed the zebrafish mef2ca mutant. When this transcription factor encoding gene is mutated, zebrafish develop dramatically variable craniofacial phenotypes. Years of selective breeding for low and high penetrance of mutant phenotypes produced strains that are either resilient or sensitive to the mef2ca mutation. Here, we compared gene expression between these strains, which revealed that selective breeding enriched for high and low mef2ca paralog expression in the low- and high-penetrance strains, respectively. We found that mef2ca paralog expression is variable in unselected wild-type zebrafish, motivating the hypothesis that heritable variation in paralog expression underlies mutant phenotype severity and variation. In support, mutagenizing the mef2ca paralogs, mef2aa , mef2b , mef2cb , and mef2d demonstrated modular buffering by paralogs. Specifically, some paralogs buffer severity while others buffer variability. We present a novel, mechanistic model for phenotypic variation where variable, vestigial paralog expression buffers development. These studies are a major step forward in understanding the mechanisms of facial variation, including how some genetically resilient individuals can overcome a deleterious mutation.
Article activity feed
-
-
Author Response
Reviewer 1
Bailon-Zambrano and colleagues were trying to answer the general question: what contributes to phenotypic variation when a gene of strong effect is mutated?
The work has several major strengths for answering this interesting question. First, they decided to study mef2ca in zebrafish for which they had previously shown that mutants displayed highly variable facial phenotypes. To learn how phenotypic variation depends on phenotypic severity, they realized they had studied more alleles, and so induced two more alleles to have three different types of molecular lesions (start codon mutation, premature stop codon, and full coding gene deletion). Investigating these alleles showed that increasingly severe alleles had more variation among individuals in the population but not necessarily more variation between …
Author Response
Reviewer 1
Bailon-Zambrano and colleagues were trying to answer the general question: what contributes to phenotypic variation when a gene of strong effect is mutated?
The work has several major strengths for answering this interesting question. First, they decided to study mef2ca in zebrafish for which they had previously shown that mutants displayed highly variable facial phenotypes. To learn how phenotypic variation depends on phenotypic severity, they realized they had studied more alleles, and so induced two more alleles to have three different types of molecular lesions (start codon mutation, premature stop codon, and full coding gene deletion). Investigating these alleles showed that increasingly severe alleles had more variation among individuals in the population but not necessarily more variation between the left and right sides of the face within individuals.
Over several years, these investigators had spent considerable effort to select lines of fish that segregate the start-codon mutation and have either severe or weak effects on facial phenotypes. wondered: what factors were selected out of the original genetic background that would increase or decrease phenotypic severity? They hypothesized that one or more of the five mef2 paralogs in zebrafish might help to ameliorate the phenotype in the low line or reciprocally intensify the phenotype in the high line. They studied expression of the mef2 paralogs in neural crest cells by single-cell transcriptomics. They found that paralogs were downregulated in the high-penetrance line with respect to an unselected line, a result expected if expression of the paralogs contributed to buffering phenotypic severity. This experiment has two weaknesses, first that the method only examined neural crest cells but we know that signals from the ectodermal and endodermal epithelia contribute to craniofacial morphologies by diffusible signals. If genes regulating craniofacial morphologies that act in epithelia had genetic variation that contributes to severity, those genes would not be investigated in these crest-only experiments. A minor problem (which is associated with the expense of the experiment) is that the scRNA-seq experiments compared only the high and unselected lines, not the low line. To address both problems, the investigators performed qPCR on RNAs extracted from whole heads of genetically mef2ca-wild types from the high and low line. In these qPCR experiments, however, they did not investigate the unselected line. Leaving out the low line in one approach and leaving out the unselected line in the other approach somewhat weakens the strength with which one can draw conclusions (e.g., the qPCR conclusion assumes that the unselected line would be intermediate between the two selected lines) but is unlikely to change the basic conclusions the authors drew. In addition, using whole heads in the qPCR experiments, while it has the advantage that it includes epithelia, does not distinguish between genes expressed only in the crest and genes expressed in other cell types, and these experiments did not test for any genes known to affect craniofacial development that are epithelium-specific.
In response to this comment, and those below, we removed the scRNA-seq comparing neural crest cells from unselected and high-penetrance strains. We replaced those data with new important results which considerably advance our model. We found significant paralog expression variation among unselected zebrafish families (Fig. 4D). These results strongly suggest that our breeding selected upon standing paralog variation the unselected parental strains. See more below.
Finally, in key experiments that are a major strength of the work and require significant effort, the researchers systematically made mutations in four of the five zebrafish mef2 paralogs (mef2aa, mef2b, mef2cb, and mef2d, all except mef2ab, which didn't become mutated despite significant effort) in the genetic background of the lowpenetrance strain and studied them in single homozygotes, in double mutants, and in various heterozygous combinations. These important experiments showed that some paralogs provided significant buffering in the low-penetrance strain, the strain that up-regulated expression of these paralogs. It would be helpful in the discussion to mention that mef2ab couldn't be mutated and a phrase added about what that means for the general conclusions - in the opinion of this reviewer, the impact of this is not great but it should be acknowledged.
We acknowledge that mef2ab couldn’t be mutated and consider what that means for the general conclusions in the text.
A strength of the experiments is that the workers quantified effects of various genotypes by focusing on the length of the symplectic, a convenient element for quantification both within single individuals and among fish in a population. It would be helpful to have a statement on the evidence that this measure is a good representative for other aspects of the phenotype.
We provide new data indicating that the symplectic cartilage length is significantly correlated with another mef2ca-associated phenotype (Fig. 1-figure supplement 2). See more below.
Finally, the paper presents a model for understanding the results presented that does a good job of summarizing the data and, importantly, suggests ways to move the analysis deeper. Missing from the description of the model is a discussion about whether the genetic variation that was selected and ultimately upregulated mef2 paralogs is in regulatory elements of the mef2 paralogs themselves or whether it might be in trans-acting transcriptional regulators that simultaneously regulate all mef2 paralogs due to the authors' hypothesized 'cryptic vestigial' functions.
We considerably revised the discussion, thoroughly considering both these possibilities.
This work is likely to have a significant impact on the fields of developmental biology, the interpretation of human mutational variation (in for example the concept of phenotypic expansion), and the way people think about the evolution of new morphologies over time. A brief comparison of the authors' results and interpretations to those of C.H. Waddington's concept of genetic assimilation would provide improved historical context and broaden the potential impact of the work.
We now include a discussion of our study in the context of Waddington’s genetic assimilation.
Reviewer 2
Bailon-Zambrano et al study the possible mechanisms that contribute to the oft-observed phenomenon that an individual mutation may be associated with variable expression of a phenotype. They focus on loss-of-function of the mef2ca gene of zebrafish, which is needed for the normal development of several craniofacial structures. They demonstrate that recessive putative loss-of-function mutant alleles of the mef2ca gene of zebrafish are associated with a range of expressivity. By focusing on one aspect of the mutant phenotype, the length of the symplectic cartilages that support the jaw, they find a correlation between the average strength of the phenotype of an allele (measured as reduction in length) and the extent of variability between mutant individuals that carry the allele. I am concerned about this conclusion and generalizations that may be drawn from focus on a single quantifiable character, the symplectic cartilage. Perhaps there is always a fixed variation in the length of this cartilage. As stronger alleles produce shorter cartilage pieces, variations in size may appear to be of greater significance when affecting shorter average length.
We now show that the symplectic cartilage length is a good proxy for other craniofacial phenotypes (Fig. 1figure supplement 2). Further, we clarify in the text that we use the coefficient of variation (standard deviation/mean) which is the accepted best practice for determining and comparing variation. We also use the F-test statistic which is the standard statistical method to test for equality of two variances. This test tells us if the standard deviations from two datasets are significantly different.
The authors hypothesize that one factor that contributes to the varied phenotypic expression of an allele (expressivity) is the co-expression of paralogs that may provide wildtype function and thus partially or wholly rescue the mutant phenotype. They test this hypothesis by "fixing" conditions where a single mutation may be expressed with low or high penetrance. By selective breeding based on phenotype, they create two sets of strains that carry an identical mef2ca mutation: one strain has high penetrance of the mutant phenotype and the other low penetrance. They then investigate the factors that are likely responsible for the high vs low penetrance. Historically we would call these factors "genetic modifiers". There is extensive literature on the nature of genetic modifiers and there are many current screens in both mice and Drosophila to identify genetic modifiers and uncover their nature, but there is little reference to these studies in the current manuscript. Further, there is previously published work that hypothesizes that one important function of paralogs in multicellular organisms is to provide a buffer to stabilize levels of gene expression needed for developmental decisions.
Following this reviewer’s suggestion, we now include many new references (increased from ~50 to >80) incorporating much of the important work leading up to our study. These include referencing both genetic modifier mutagenesis screens, paralogous buffering in other systems, and “natural” modifier studies that set the stage for our work.
The authors find that paralogs of the mef2ca gene are expressed in cells that normally express mef2ca, and that these paralogs are expressed at higher levels in the mutant strain with low penetrance than in the mutant strain with high penetrance. They say that selection for high penetrance of the mef2ca mutant phenotype "leads to down-regulation" of paralog expression. As the authors only show that paralog expression is at lower levels in high penetrance vs low penetrance strains, it is not clear what they mean by "down-regulation". Perhaps their breeding scheme has only "captured" what is natural variation and there is no active mechanism of "down-regulation". The authors need to clarify what they mean.
Thank you for this suggestion. We clarified that we do not mean active down or up regulation but rather selection on preexisting genetic variation. This conclusion is supported by new data (Fig. 4D).
The authors also find that individuals from the high penetrance strains that don't carry the mef2ca mutation (they are wildtype for this gene) sometimes exhibit mef2ca mutant characters. They suggest the reduced paralog expression is responsible for the occasional emergence of the mef2ca mutant characters. In contrast with this suggestion, the authors later claim the paralogs "have no function" in craniofacial development. The authors need to clarify their thoughts about what is paralog function in craniofacial development and why reduced paralog function might contribute to the expression of mef2ca mutant characters. This topic is worthy of discussion.
We considerably revised our discussion of this topic including our interpretation that the decreased expression of mef2ca in high penetrance strain led to the phenotypes we observe in mef2ca wild types from this strain. We also are more careful with our language, stating that the paralog mutants are indistinguishable from wild types, rather than stating that paralogs do not function in craniofacial development. In fact, they do function in craniofacial development, as buffers. Thank you for this suggestion that strengthened our manuscript.
The authors claim is there is both up-regulation of paralogs in low penetrance strains and down-regulation of paralogs in high penetrance strains. As they only compare steady state levels of expression in each strain, they can only reasonably conclude that there are differences - they seem to imply a mechanism and they need to be clear about what they are thinking.
Excellent point. In the revised manuscript, we are clear that there is not active up or down regulation but rather selection upon preexisting variation.
They hypothesize that paralog expression in the low penetrance strain masks the effects of loss of mef2ca. They test this by creating CRISPR-engineered mutations of two paralogs and examining the effects of the paralog mutations in wildtype fish or in fish carrying the mef2ca mutation. They find the putative loss-offunction mutations in the paralogs have no effect in wildtype backgrounds and conclude these paralog genes have no function in craniofacial development. However, the paralog mutations enhance the mutant phenotype in fish that carry the mef2ca mutation. This provides strong evidence consistent with the model that the elevated expression of the paralogs functions to reduce the severity of the phenotype associated with the mef2ca mutation.
Reviewer 3
In this elegant genetic study, Bailon-Zambrano et al. draw on classical genetic concepts to address the clinically pertinent question of how genetic variants in the same gene can yield wildly different phenotypes in different individuals. They focus on the Mef2c gene, which is required for craniofacial and cardiac development in humans and model organisms yet shows highly variable phenotypes across and within individuals. Previous work by this lab had established that zebrafish mef2ca craniofacial phenotypes are highly variable and, importantly, that this variability is heritable and can be selectively bred for low vs. high penetrance. The authors hypothesize that vestigial expression of paralogous genes variably compensates for loss of mef2ca, such that individuals with higher levels of paralogous genes will show lessened severity and vice versa. To test their hypothesis, they methodically quantify the penetrance, expressivity, and variability of all known mef2caassociated craniofacial phenotypes in fish carrying 1) different mef2ca mutations, 2) the same mutation but after selecting for high vs. low penetrance for many generations, and 3) mef2ca mutations combined with mutations in paralogous genes. They find that not only does allele severity directly correlate with variation, but also that different paralogs buffer the severity and variability of different craniofacial phenotypes. Another particularly interesting finding is that some of the craniofacial phenotypes are apparent even in mef2ca wildtypes from the high penetrance strain, which they explain by the very low expression of paralogs on this background. A weakness of the study is that the authors do not directly show whether paralog expression is increased in the low-penetrance strain relative to the initial, unselected genetic background. It is therefore not clear whether the selection for low penetrance worked in this manner, as the authors imply. Overall, the authors have achieved an important step forward in understanding the genetic basis for the high variability of human faces among both healthy individuals and those with craniofacial anomalies.
We can’t go back (over ten generations) to survey the original parental strain. However, we can use the unselected AB strain as a proxy for the initial unselected genetic background. In an important addition to the manuscript, we found significant paralog expression variation between unselected AB families (Fig. 4D). These results strongly suggesting there is cryptic, standing paralog expression variation that we selected upon. We would like to thank the reviewer for this excellent critique which motivated these important new experiments considerably advancing our model.
-
Evaluation Summary:
In this elegant genetic study, Bailon-Zambrano and colleagues draw on classical genetic concepts to address the clinically pertinent question of how genetic variants in the same gene can yield wildly different phenotypes in different individuals. From their case study they conclude that a major contributor is variation in paralog expression. The question addressed is of great interest to evolutionary and developmental biologists in general and to those studying the evolution of developmental mechanisms in particular.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript.The reviewers remained anonymous to the authors.)
-
Reviewer #1 (Public Review):
Bailon-Zambrano and colleagues were trying to answer the general question: what contributes to phenotypic variation when a gene of strong effect is mutated?
The work has several major strengths for answering this interesting question. First, they decided to study mef2ca in zebrafish for which they had previously shown that mutants displayed highly variable facial phenotypes. To learn how phenotypic variation depends on phenotypic severity, they realized they had studied more alleles, and so induced two more alleles to have three different types of molecular lesions (start codon mutation, premature stop codon, and full coding gene deletion). Investigating these alleles showed that increasingly severe alleles had more variation among individuals in the population but not necessarily more variation between the …
Reviewer #1 (Public Review):
Bailon-Zambrano and colleagues were trying to answer the general question: what contributes to phenotypic variation when a gene of strong effect is mutated?
The work has several major strengths for answering this interesting question. First, they decided to study mef2ca in zebrafish for which they had previously shown that mutants displayed highly variable facial phenotypes. To learn how phenotypic variation depends on phenotypic severity, they realized they had studied more alleles, and so induced two more alleles to have three different types of molecular lesions (start codon mutation, premature stop codon, and full coding gene deletion). Investigating these alleles showed that increasingly severe alleles had more variation among individuals in the population but not necessarily more variation between the left and right sides of the face within individuals.
Over several years, these investigators had spent considerable effort to select lines of fish that segregate the start-codon mutation and have either severe or weak effects on facial phenotypes. They wondered: what factors were selected out of the original genetic background that would increase or decrease phenotypic severity? They hypothesized that one or more of the five mef2 paralogs in zebrafish might help to ameliorate the phenotype in the low line or reciprocally intensify the phenotype in the high line. They studied expression of the mef2 paralogs in neural crest cells by single-cell transcriptomics. They found that paralogs were downregulated in the high-penetrance line with respect to an unselected line, a result expected if expression of the paralogs contributed to buffering phenotypic severity. This experiment has two weaknesses, first that the method only examined neural crest cells but we know that signals from the ectodermal and endodermal epithelia contribute to craniofacial morphologies by diffusible signals. If genes regulating craniofacial morphologies that act in epithelia had genetic variation that contributes to severity, those genes would not be investigated in these crest-only experiments. A minor problem (which is associated with the expense of the experiment) is that the scRNA-seq experiments compared only the high and unselected lines, not the low line. To address both problems, the investigators performed qPCR on RNAs extracted from whole heads of genetically mef2ca-wild types from the high and low line. In these qPCR experiments, however, they did not investigate the unselected line. Leaving out the low line in one approach and leaving out the unselected line in the other approach somewhat weakens the strength with which one can draw conclusions (e.g., the qPCR conclusion assumes that the unselected line would be intermediate between the two selected lines) but is unlikely to change the basic conclusions the authors drew. In addition, using whole heads in the qPCR experiments, while it has the advantage that it includes epithelia, does not distinguish between genes expressed only in the crest and genes expressed in other cell types, and these experiments did not test for any genes known to affect craniofacial development that are epithelium-specific.
Finally, in key experiments that are a major strength of the work and require significant effort, the researchers systematically made mutations in four of the five zebrafish mef2 paralogs (mef2aa, mef2b, mef2cb, and mef2d, all except mef2ab, which didn't become mutated despite significant effort) in the genetic background of the low-penetrance strain and studied them in single homozygotes, in double mutants, and in various heterozygous combinations. These important experiments showed that some paralogs provided significant buffering in the low-penetrance strain, the strain that up-regulated expression of these paralogs. It would be helpful in the discussion to mention that mef2ab couldn't be mutated and a phrase added about what that means for the general conclusions - in the opinion of this reviewer, the impact of this is not great but it should be acknowledged.
A strength of the experiments is that the workers quantified effects of various genotypes by focusing on the length of the symplectic, a convenient element for quantification both within single individuals and among fish in a population. It would be helpful to have a statement on the evidence that this measure is a good representative for other aspects of the phenotype.
Finally, the paper presents a model for understanding the results presented that does a good job of summarizing the data and, importantly, suggests ways to move the analysis deeper. Missing from the description of the model is a discussion about whether the genetic variation that was selected and ultimately upregulated mef2 paralogs is in regulatory elements of the mef2 paralogs themselves or whether it might be in trans-acting transcriptional regulators that simultaneously regulate all mef2 paralogs due to the authors' hypothesized 'cryptic vestigial' functions.
This work is likely to have a significant impact on the fields of developmental biology, the interpretation of human mutational variation (in for example the concept of phenotypic expansion), and the way people think about the evolution of new morphologies over time. A brief comparison of the authors' results and interpretations to those of C.H. Waddington's concept of genetic assimilation would provide improved historical context and broaden the potential impact of the work.
-
Reviewer #2 (Public Review):
Bailon-Zambrano et al study the possible mechanisms that contribute to the oft-observed phenomenon that an individual mutation may be associated with variable expression of a phenotype. They focus on loss-of-function of the mef2ca gene of zebrafish, which is needed for the normal development of several craniofacial structures. They demonstrate that recessive putative loss-of-function mutant alleles of the mef2ca gene of zebrafish are associated with a range of expressivity. By focusing on one aspect of the mutant phenotype, the length of the symplectic cartilages that support the jaw, they find a correlation between the average strength of the phenotype of an allele (measured as reduction in length) and the extent of variability between mutant individuals that carry the allele. I am concerned about this …
Reviewer #2 (Public Review):
Bailon-Zambrano et al study the possible mechanisms that contribute to the oft-observed phenomenon that an individual mutation may be associated with variable expression of a phenotype. They focus on loss-of-function of the mef2ca gene of zebrafish, which is needed for the normal development of several craniofacial structures. They demonstrate that recessive putative loss-of-function mutant alleles of the mef2ca gene of zebrafish are associated with a range of expressivity. By focusing on one aspect of the mutant phenotype, the length of the symplectic cartilages that support the jaw, they find a correlation between the average strength of the phenotype of an allele (measured as reduction in length) and the extent of variability between mutant individuals that carry the allele. I am concerned about this conclusion and generalizations that may be drawn from focus on a single quantifiable character, the symplectic cartilage. Perhaps there is always a fixed variation in the length of this cartilage. As stronger alleles produce shorter cartilage pieces, variations in size may appear to be of greater significance when affecting shorter average length.
The authors hypothesize that one factor that contributes to the varied phenotypic expression of an allele (expressivity) is the co-expression of paralogs that may provide wildtype function and thus partially or wholly rescue the mutant phenotype. They test this hypothesis by "fixing" conditions where a single mutation may be expressed with low or high penetrance. By selective breeding based on phenotype, they create two sets of strains that carry an identical mef2ca mutation: one strain has high penetrance of the mutant phenotype and the other low penetrance. They then investigate the factors that are likely responsible for the high vs low penetrance. Historically we would call these factors "genetic modifiers". There is extensive literature on the nature of genetic modifiers and there are many current screens in both mice and Drosophila to identify genetic modifiers and uncover their nature, but there is little reference to these studies in the current manuscript. Further, there is previously published work that hypothesizes that one important function of paralogs in multicellular organisms is to provide a buffer to stabilize levels of gene expression needed for developmental decisions.
The authors find that paralogs of the mef2ca gene are expressed in cells that normally express mef2ca, and that these paralogs are expressed at higher levels in the mutant strain with low penetrance than in the mutant strain with high penetrance. They say that selection for high penetrance of the mef2ca mutant phenotype "leads to down-regulation" of paralog expression. As the authors only show that paralog expression is at lower levels in high penetrance vs low penetrance strains, it is not clear what they mean by "down-regulation". Perhaps their breeding scheme has only "captured" what is natural variation and there is no active mechanism of "down-regulation". The authors need to clarify what they mean.
The authors also find that individuals from the high penetrance strains that don't carry the mef2ca mutation (they are wildtype for this gene) sometimes exhibit mef2ca mutant characters. They suggest the reduced paralog expression is responsible for the occasional emergence of the mef2ca mutant characters. In contrast with this suggestion, the authors later claim the paralogs "have no function" in craniofacial development. The authors need to clarify their thoughts about what is paralog function in craniofacial development and why reduced paralog function might contribute to the expression of mef2ca mutant characters. This topic is worthy of discussion.
The authors claim is there is both up-regulation of paralogs in low penetrance strains and down-regulation of paralogs in high penetrance strains. As they only compare steady state levels of expression in each strain, they can only reasonably conclude that there are differences - they seem to imply a mechanism and they need to be clear about what they are thinking.
They hypothesize that paralog expression in the low penetrance strain masks the effects of loss of mef2ca. They test this by creating CRISPR-engineered mutations of two paralogs and examining the effects of the paralog mutations in wildtype fish or in fish carrying the mef2ca mutation. They find the putative loss-of-function mutations in the paralogs have no effect in wildtype backgrounds and conclude these paralog genes have no function in craniofacial development. However, the paralog mutations enhance the mutant phenotype in fish that carry the mef2ca mutation. This provides strong evidence consistent with the model that the elevated expression of the paralogs functions to reduce the severity of the phenotype associated with the mef2ca mutation.
-
Reviewer #3 (Public Review):
In this elegant genetic study, Bailon-Zambrano et al. draw on classical genetic concepts to address the clinically pertinent question of how genetic variants in the same gene can yield wildly different phenotypes in different individuals. They focus on the Mef2c gene, which is required for craniofacial and cardiac development in humans and model organisms yet shows highly variable phenotypes across and within individuals. Previous work by this lab had established that zebrafish mef2ca craniofacial phenotypes are highly variable and, importantly, that this variability is heritable and can be selectively bred for low vs. high penetrance. The authors hypothesize that vestigial expression of paralogous genes variably compensates for loss of mef2ca, such that individuals with higher levels of paralogous genes …
Reviewer #3 (Public Review):
In this elegant genetic study, Bailon-Zambrano et al. draw on classical genetic concepts to address the clinically pertinent question of how genetic variants in the same gene can yield wildly different phenotypes in different individuals. They focus on the Mef2c gene, which is required for craniofacial and cardiac development in humans and model organisms yet shows highly variable phenotypes across and within individuals. Previous work by this lab had established that zebrafish mef2ca craniofacial phenotypes are highly variable and, importantly, that this variability is heritable and can be selectively bred for low vs. high penetrance. The authors hypothesize that vestigial expression of paralogous genes variably compensates for loss of mef2ca, such that individuals with higher levels of paralogous genes will show lessened severity and vice versa. To test their hypothesis, they methodically quantify the penetrance, expressivity, and variability of all known mef2ca-associated craniofacial phenotypes in fish carrying 1) different mef2ca mutations, 2) the same mutation but after selecting for high vs. low penetrance for many generations, and 3) mef2ca mutations combined with mutations in paralogous genes. They find that not only does allele severity directly correlate with variation, but also that different paralogs buffer the severity and variability of different craniofacial phenotypes. Another particularly interesting finding is that some of the craniofacial phenotypes are apparent even in mef2ca wild-types from the high penetrance strain, which they explain by the very low expression of paralogs on this background. A weakness of the study is that the authors do not directly show whether paralog expression is increased in the low-penetrance strain relative to the initial, unselected genetic background. It is therefore not clear whether the selection for low penetrance worked in this manner, as the authors imply. Overall, the authors have achieved an important step forward in understanding the genetic basis for the high variability of human faces among both healthy individuals and those with craniofacial anomalies.
-
