|
Author
|
Topic: Duplicate Genes Implications
|
Josh
Member
Member # 405
|
posted 07. January 2003 14:01
I wanted to stimulate some discussion regarding a recent article (and corresponding News and Views) in Nature. Reference Gu et al., Nature 421, 63-66, and 31-32. Following excerpts (first two paragraphs followed by the final paragraph of both articles) highlight the topic:
Original Article:
"Deleting a gene in an organism often has little phenotypic effect, owing to two mechanisms of compensation. The first is the existence of duplicate genes: that is, the loss of function in one copy can be compensated by the other copy or copies. The second mechanism of compensation stems from alternative metabolic pathways, regulatory networks, and so on. The relative importance of the two mechanisms has not been investigated except for a limited study, which suggested that the role of duplicate genes in compensation is negligible. The availability of fitness data for a nearly complete set of single-gene-deletion mutants of the Saccharomyces cerevisiae genome has enabled us to carry out a genome-wide evaluation of the role of duplicate genes in genetic robustness against null mutations. Here we show that there is a significantly higher probability of functional compensation for a duplicate gene than for a singleton, a high correlation between the frequency of compensation and the sequence similarity of two duplicates, and a higher probability of a severe fitness effect when the duplicate copy that is more highly expressed is deleted. We estimate that in S. cerevisiae at least a quarter of those gene deletions that have no phenotype are compensated by duplicate genes.
No correlation was found between the sequence similarity of duplicate genes and the fitness effect of a null mutation in one of the two duplicates when functional data from the yeast S. cerevisiae was analysed previously10. It was therefore concluded that gene duplications contribute little to the ability of an organism to withstand mutations (genetic robustness), although they may be responsible for a small fraction of weak, null-mutation phenotypes12. Because this conclusion was based on only 45 duplicate genes, however, the issue deserves further investigation. Indeed, this conclusion is not supported by a limited analysis of a third of the genes in the yeast genome1 and is contrary to the general observation of relaxed selective constraints after gene duplication.
Although our estimates are compatible with the view that interactions among unrelated genes rather than duplicate genes are the main cause of genetic robustness against mutations10, 18, two additional factors need to be considered. First, because we have considered only five growth conditions, it is possible that when a gene deletion showed no effect in any of these conditions it was not due to compensation by other genes but was because the gene deleted was not related to the growth conditions used. Intuitively, when more growth conditions are studied, both the proportion of duplicate genes and the proportion of singletons that show only a weak or no effect of deletion on growth rate will decrease. Indeed, the two proportions were 70.9% and 49.2% when only the YPD growth condition was considered (data not shown), but became 64.3% and 39.5% when the five growth conditions shown in Fig. 1a were used. The decrease is larger for singletons than for duplicate genes, probably because duplicate genes have on average a stronger overlap in function than do singletons and so can compensate each other in a wider range of conditions. For this reason, our lower bound of 23% for the relative contribution of duplicate genes to compensation for null mutations is likely to be an underestimate. Second, a singleton in this or other studies could actually have one or more paralogues in the genome that cannot be detected by the criteria used but still overlap in function. Thus, gene duplication might be the ultimate origin of functional compensation for some 'singletons'. In conclusion, whether the contribution of gene duplication to genetic robustness is really less important than interactions among unrelated genes is an issue that remains to be resolved by further studies."
News and Views:
"Duplicated genes are common in genomes, perhaps because they provide redundancy: if one copy is inactivated, the other can still work. A new study quantifies the effects of deleting 'singletons' and duplicated genes in yeast.
In fairy tales, things frequently come in twos: there are, for instance, two witches ruling over different parts of the land of Oz, two ugly sisters vying for the attention of Cinderella's prince, and so on and so on. And the phenomenon of duplication is not restricted to stories. In eukaryotes (loosely speaking, those organisms, such as humans, whose DNA is packaged into cell nuclei), genomes seem to be far from optimally designed, in that most stretches of DNA sequence do not code for proteins, and even those small portions that do are often duplicated. Why do organisms tolerate such apparent wastage? Gu and colleagues1 tackle this question on page 63 of this issue, looking specifically at the effects of duplicated genes on the 'fitness' of individuals.
An important line of thinking about why duplicated genes might arise goes back 30 years to Susumo Ohno2, who stated that "natural selection merely modified while redundancy created". Ohno reasoned that gene (and even genome) duplications are not a burden on the organism, but rather the raw material for evolutionary diversification — in other words, duplication allows new gene functions to evolve. One copy of a gene can carry out the original task while the duplicate becomes free to accumulate mutations, possibly developing new functions and allowing the big steps in evolution to occur. In today's era of wholesale genome sequencing, Ohno's hypothesis has gained many new adherents through the recognition that duplicate genes are abundant in most genomes and that significant portions of genomes are repeated. But, in general, the actual effects of 'singletons' and duplicated genes on evolutionary fitness — that is, on roughly how well different individuals fare compared with others in terms of reproduction — have not been well studied at the whole-genome level.
On the other side of the coin, gene duplicates appear to have another important function: they can buffer the genome against environmental perturbations and mutations, because if one copy of the gene is somehow inactivated, another with the same or a similar function can be used instead. Such genetic redundancy is a headache for researchers trying to determine the role of a particular gene, because the standard technique of knocking out that gene in an organism might not have a noticeable effect, thanks to functional substitution by the duplicate. Gu et al.1 shed new light on this issue.
We are only now beginning to comprehend just how malleable genomes are, and also how resilient they are in the face of so much genetic perturbation; for instance, rearrangements and duplications of chromosomal segments are also commonplace8, 9. Gu et al.1 have provided the first estimate (23–59%) of the contribution of duplicated genes to genetic robustness. This may be one reason why duplicated genes do not diverge to produce pseudogenes, or 'die', as quickly or as often as had been predicted on the basis of population-genetics theory10. I would guess that the existence of multiple gene functions and their recruitment into novel gene networks provide another explanation. But more needs to be learned about the evolution of gene networks, through comparisons of complete genome sequences and through further functional-genomic analyses, before this question can be answered."
So a few offhand questions:
Can we really assess the genome and proclaim that it is filled with junk (and not optimally designed) in light of this study?
Does the fact that duplicated genes contribute to genetic robustness lead to the conclusion of design or evolution?
Taking this data into account does the presence of duplicated genes by themselves indicate clearly that they must be products of evolution?
What does the fact that deleting one copy of a duplicate gene 12.4% of the time in S. cerevisiae causes leathality indicate about the purpose of duplicate genes? Simply raw material for evolution, or perhaps requisite components of a well-designed system needed for reasons not yet determined? Consider also the comments concerning the growth conditions that the authors make in the last paragraph of the original paper in regards to this question. The comments concerning growth conditions directly relate to arguments I tried to make earlier to Argon and others who claim that artificial lab environments are in fact suitable for predicting general features of evolution/ adaptability. (see "Does Darwinism Predict the Absence of Irreducible Complexity?" http://www.iscid.org/ubbcgi/ultimatebb.cgi?ubb=get_topic;f=6;t=000165;p=3) This data proves that not only are culture conditions relevant, they are imperative and change the data outcome depending on the conditions. YPD vs. minimal media will give different gene requirements, hence the argument I was making on this old thread.
I thought this recent article was pretty interesting and look forward to hearing feedback /viewpoints.
[ 07. January 2003, 14:02: Message edited by: Josh ]
IP: Logged
|
|
Art
Member
Member # 179
|
posted 07. January 2003 23:31
Hi Josh,
About yur questions:
quote:
So a few offhand questions:
Can we really assess the genome and proclaim that it is filled with junk (and not optimally designed) in light of this study?
Yup.
"Junk DNA" and expressed gene families are two different things. I don't see the connection you seem to be making.
quote:
Does the fact that duplicated genes contribute to genetic robustness lead to the conclusion of design or evolution?
The connection between "evolution" and gene duplication seem obvious to me.
OTOH, the nature of gene families seems far too haphazard to read much by way of design into the matter.
quote:
Taking this data into account does the presence of duplicated genes by themselves indicate clearly that they must be products of evolution?
No offense, but this question is a little confusing, and it gets more so every time I re-read it.
What are you asking?
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 08. January 2003 02:12
This is a fascinating topic that I've been thinking about for some time: the evoluion of redundancy. My structural engineer friend Ryan Huxley and I have been kicking around ideas related to the evolution of redundancy for some time; we even discussed a few of our ideas here:
http://www.iscid.org/ubbcgi/ultimatebb.cgi?ubb=get_topic;f=6;t=000135
In light of that earlier discussion, I've had some questions rattling around in my brain for awhile:
1. It seems that redundancy is a characteristic of designed systems where the level of complexity and the need for robust function are very high. We see this in the space shuttle, for example, in crucial life-support systems that simply cannot fail. Furthermore, these redundant, backup systems provide no initial benefit--they are installed entirely with a future contingency "in mind". However, evolution has no "mind" or goal (which in this case is the robust functioning of a complex system) that it can work towards. Since a redundant system provides no initial benefit, it's difficult to see how it could evolve without intelligent input. This conceptual problem becomes more acute as the complexity of the backup, redundant system increases, since an unguided process would have to produce a system that is effectively "hidden" from immediate use and is basically useless except in the rare case of failure of the primary system. Furthermore, the construction and maintainence of this backup system will entail a fitness cost on the organism that carries it, which might counteract any selective advantage to be gained by having such a system. So how does a redundant system evolve?
2. The most potent counter-argument I have heard is that redundancy is predicted by evolutionary processes! The idea is that a redundant system, created by genetic duplication, is free to evolve freely without destructive consequences for the organism, and hence new functionality can be generated. Thus, far from being a challenge for evolution, redundancy is in fact a prediction and confirmation of evolutionary processes.
3. However, this argument seems a bit superficial to me. Doesn't redundancy actually promote stability, not change? After all, one point made in the Gu et al. paper is that genetic redunancy makes research hard because point mutations simply don't have an effect on the system--they don't "knock out" functionality because another gene can compensate. If this is the case, why should we expect that evolution can take advantage of this redundancy and generate novelty? How many mutations have to accumulate in one redundant system, before it acquires new function? Obviously, these changes cannot occur under positive selective pressure because they are not "seen" by selection (they are masked by the other, compensating genetic system). Therefore these changes have to occur by neutral evolution; neutral mutations have to fortunately "sweep" the population and be preserved. As any good population geneticist will tell you, the probability that a neutral mutation will be preserved in a population is the same as the rate of neutral mutations arising in that population--generally around one in a million (per gene) or less. If there are more than a couple of mutations that have to accrue this way, the probabilities quickly become prohibitively low.
4. As I mentioned earlier, redundancy is used in engineering to achieve the goal of robustness. Could this be the case in biological systems? Could the systems be complex enough that a single copy of some genes is simply not enough to ensure robust function over the lifetime of an individual? We know that we humans need two copies of the retinoblastoma gene (RB) because if we are born with one copy mutated, the chances of acquiring cancer of the retina (before the age of 20 or so) is nearly 100%. However, most of us have two normal copies of the gene, and most of us never get cancer of the retina. Apparently, robustness of the system requires redundancy of function, at least in this biological example. Could this be the case in other situations? Is this the reason why so many redundant genes seem to still be under selective pressure?
In particular, I think it would be worthwhile to explore the question of whether redundancy in biology is a challenge for evolution, or is in fact a prediction and verification of evolutionary theory. How can we tell?
John
[ 08. January 2003, 02:13: Message edited by: John Bracht ]
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 08. January 2003 11:11
John Bracht asks:
quote:
In particular, I think it would be worthwhile to explore the question of whether redundancy in biology is a challenge for evolution, or is in fact a prediction and verification of evolutionary theory. How can we tell?
I think that redundancy, at least with regard to duplication of genes and the generation of gene families is a verification of common descent with modification. After all, there are other ways of producing robustness without replicating gene families.
Duplications of genes can alter the expression (timing, location & amount) of their products and thus are not necessarily neutral to selection. They can also back up essential functions that could otherwise be lost. One example is the case of retinoblastoma mentioned by John Bracht. Now it is true that multiple copies (actually, variants) of similar genes can act conservatively to retard genetic change, such as through recombination between similar sequences. This, in itself will have a selective advantage in some cases, permitting repair of selectively negative mutations. Then again, having redundancy can permit one of the copies to 'explore' alternative roles, and thus could provide the material for future variation. But there is nothing that makes these two apparently opposite possibilities a permanent, 'either/or' situation in any particular case. Conditions could change (e.g. environmental variation, interaction with different combinations of genes that arise from reproduction, acquisition of positive mutations in one copy & etc.), that shift the balance between conservative and diversifying forces over time. I think the ranges of these effects are readily apparent when we look at the levels of variation exhibited by duplicated genes and gene families in organisms. Some show tight conservation, others far less (Try using keywords searches with terms like 'duplication', 'gene family' & 'evolution' in Medline for examples. The question of how redundancy may arise in biology is being actively discussed and evaluated in the literature*).
What is clear is that one cannot make sweeping generalizations about how duplications will change over time, whether leading to homogeneity or heterogeneity -- Examples can only be evaluated on a case-by-case basis.
I suppose this is another case where chemostats could have an application. For example, some duplications are unstable in bacteria but can be held in place through growth on selective media. These are often nutrient sources (or antibiotics) for which the bacteria need high levels of expression to accommodate. In a chemostat, if one cycled different media that was selective and non-selective for a particular duplication, one would be able to see how cells adapted.
Added notes:
I just remembered: The story behind the 'adaptive mutation' of the lac operon in E. coli (by Hall and others) turned out to be a case of gene duplication and mutation, instead of 'directed mutation'. At least in this case of redundancy, an evolutionary mechanism seems to have been confirmed.
* In the thread referenced by John, one correspondent (fish) cited a few papers that may be relevant to John's questions.
[ 08. January 2003, 11:57: Message edited by: Argon ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 08. January 2003 12:57
John wrote quote:
In particular, I think it would be worthwhile to explore the question of whether redundancy in biology is a challenge for evolution, or is in fact a prediction and verification of evolutionary theory. How can we tell?
There's a possibility, of course, that's it's neither a critical prediction nor a challenge. I'm reminded (by the space shuttle example) of the employment of Newtonian mechanics to produce gravitational "boosts" for far-traveling spacecraft. That's neither a prediction of Newtonian mechanics (before the technique was thunk up) nor a challenge: It's merely a consequence that wasn't thought of until hundreds of years after Newton. It is consistent with Newtonian mechanics and is an application of it. (I'm not super sure that it's an approriate analogy, either, but I kind of like it. ![[Smile]](smile.gif) )
One further thought was stimulated by the space shuttle example. In the 1960s I worked in aerospace and defense on the design and evaluation of control systems for various vehicles including early versions of the Polaris IRBM and the Apollo Command Module. Redundancy in those days, even in the manned Apollo, was employed less than it is in the space shuttle, primarily because of technological limitations. Even electronic components in those days were (relatively) large and heavy and there were weight and space limits on how much redundancy could incorporated and still get the bird off the ground. One can't judge the general utility and employment of redundancy, in human-designed objects or in nature, from the current state of a system that has had a long history of (designed or evolved) development. External constraints can intervene and distort the picture.
Redundant systems are not free - they require resources and thus incur costs. The evolutionary process is demonstrably an algorithm capable of weighing and reconciling those kinds of cost/benefit problems. So, at least to some extent, are human designers. Paul Layzell's "Populational Fault Tolerance" notion (see his dissertation here) is another way of achieving robustness in both biological and human-designed systems. There the "redundancy" and consequent robustness is at the population level, not the individual organism/entity level. (That, btw, is a way that analogies with human design can lead one astray: thinking in terms of variability in populations in biological systems versus thinking in terms of an individual entity and analysis of a single homogenous design in human design.)
If redundancy is a common feature of more critical (in some well-defined sense) biological subsystems, it seems to me that it could follow from either design principles derived from observation of human artifacts or from evolutionary principles, and thus doesn't distinguish between them.
RBH
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 08. January 2003 17:58
This topic interests me for a quite practical reason: My firm uses EAs to evolve systems that have to operate in the real world in the face of varying and unpredictable vicissitudes, and their robustness is a concern of ours. So I'd like it to proceed if at all possible.
If this discussion is to go anywhere, I suggest we first establish a common language and some core references. A lot of wheels have already been invented in this general domain - a Google search on "robustness redundancy evolution" (the union of the three words, not the three-word phrase) yielded 7,480 hits.
A first distinction we might make is between "redundancy" and "degeneracy." This paper is a good place to start: Measures of degeneracy and redundancy in biological networks.
"Robustness" is a generic term that has no specific technical meaning that I'm aware of. In general it implies that a system is stable in the face of perturbations or insults, but I know of no way to quantify it generally, in a way that is system-independent, at least not in the kinds of systems we're concerned with here. Anyone else?
RBH
IP: Logged
|
|
Ryan Huxley
Member
Member # 366
|
posted 09. January 2003 17:30
I'm a bit limited in time, so this will be a rather brief response. However, I hope to respond to some of the other points made in some of the previous posts in the near future (next few days?...).
In the last post by RBH, he was interested in defining terms to help promote the discussion, not to mention other ways to consider redundancy from a design perspective.
First, as mentioned by John Bracht, redundancy is recognized in the engineering fields as a desired aspect of design. In fact, in structural engineering, redundancy concepts are actually incorporated in the building code (I can provide references for those interested). For buildings that suffered damage during seismic events, lack of redundancy is often sited as at least one of the contributing factors to poor performance. At least in the structural engineering field (and likely other engineering fields), incorporating redundancy into a design is considered good engineering.
As for definition of terms, there are mathematical expressions relating redundancy and reliability. Very loosely, reliability may be thought of as (1 - probability of failure). There are general mathematical expressions, at least in the structural engineering literature, that can show how increasing redundancy increases reliability. This makes sense intuitively. However, whenever there is energy involved, I think it becomes a bit trickier to make that generalization.
As for redundancy in the genome, how it arose in the first place is what is interesting. I think that once redundancy is present, it will be difficult to determine whether design or evolution may be credited. Though, population genetics may provide insight and a means of differentiating between designed redundancy versus evoloved redundancy, as John Bracht alluded to.
As for junk DNA, I think it may be a bit soon to claim that this supposedly functionless DNA is truly functionless. There are examples (I know, I need to provide some references here - maybe later) of deletion of introns (the junk part - I'll admit that since this is not my field, I may be getting introns and exons mixed up) leads to problems (in some cases, cancer developed). Additionally, they've found that when reading a given sequence in one direction, it's an intron, but in the other direction, it's an exon. I think it's a bit premature to claim that we understand the entire genome and the majority of it is junk when examples like are beginning to arise more and more often.
IP: Logged
|
|
|
Printer-friendly view of this topic