3 ms·

I'm going to go against this and say no. We swap gene locations, all the time wholesale in mice, and usually the effect is limited. That doesn't get you a pap

I'm going to go against this and say no. We swap gene locations, all the time wholesale in mice, and usually the effect is limited. That doesn't get you a paper, so of course you publish when something... more interesting crops up. It's just that there are many well-studied systems where where there are profound and severe effects that are singly extremely consequential, where you go "huh, wouldn't have guessed that" and the cause turns out to be really subtle. If we're going with eukarya writ large, it's even less. Basically entire programs of plant husbandry is the industrial process of performing chromosomal translocation and overloading, and picking the one out a thousand? ten thousand? that reveals a commercially useful variant. And yeast, well we've sliced and diced those chromosomes like no one's business (but it turns out they are particularly robust). I would go so far to say as "the general rule is that chromosomal location doesn't matter in yeast, with a handful of exceptions". Conversely, there is one gene whose chromosomal location is basically invariant in prokarya, that's dnaA, which pretty much "has to be close to the ori".

jballanc 7y ago
So, two quick points in response: 1. Yeah, I probably overstated the case. It's not as if it's impossible to move a gene somewhere else on a chromosome and have it function completely normally. As you point out, we do this all the time. That said... 2. Genetics is hard, and I'm of the school of thought that one reason it is so hard is because systems are more fragile in meaningful ways than we assume. That is to say, if you attempt a gene translocation 100 times and get 3 successes, most geneticists would be happy with that result and not question the 97 failures. And, indeed, it is entirely possible that the quality of the reagents, the stringency of the protocols, or any of an infinite number of other variables that have little or nothing to do with genetics could be the cause of those 97 failures...or, it could be that 97 times the translocation worked but the product wasn't viable. I know from experience that it is very hard to distinguish between these possibilities, and furthermore that in the publish-or-perish world of academia today there's not any motivation for navel gazing into the reason most genetic protocols are not more efficient. That said, I've seen hints in my past work that make me think location, gene dose, DNA secondary and tertiary structure, etc. are more consequential than the common wisdom would have you believe. (Oh, and for anyone still reading, I'll share the brain teaser that was my light-bulb moment for the extent of gene dose effect in prokarya: it takes 40 min to replicate the E. coli genome, but during exponential growth E. coli can double in number every 20 min. How?)
sjg007 7y ago
https://www.annualreviews.org/doi/abs/10.1146/annurev-genet-112414-055008 https://www.annualreviews.org/doi/abs/10.1146/annurev-genet-... and https://www.sciencedaily.com/releases/2010/10/101014144312.htm https://www.sciencedaily.com/releases/2010/10/101014144312.h...
- dnautics 7y ago
  The second one relates to a feature at the metaorganismal level; namely the evolution rate. Fun fact: human chromosome 1 is the merger of two simian chomosomes and has two centromeres
gravelc 7y ago
I'm going to disagree with go with 'sometimes yes and sometimes no'. Take a construct and transform it into a plant. Check the expression of the transgene from hundreds of different transformation events, and there'll be huge variation, from highly expressed to silenced, even though all inserted transgenes have the same promoter and terminator. Location matters. However, if the function of gene A is redundant, or it goes to a location when expression is not affected, or it's just not a very important gene, then it probably doesn't have any great phenotypic effect.
- dnautics 7y ago
  I would make a bet. Randomly generate 100 spots in the mouse genome, let's say maybe not in centromeric locations. Don't correct for being inside an ORF or a promoter. Pick 100 random translocations swapping these spots. 95 of them will be viable with no obvious phenotype.
  - hyperbovine 7y ago
    In how many samples are you doing this? It’s all about effect size.
  - klmr 7y ago
    I broadly agree with your point but careful when saying “no obvious phenotype”. As my former boss likes to say about work in C. elegans: “Worms can’t talk. They seem healthy but maybe they’re extremely unhappy.” Less flippantly, we know that biological systems in general have huge plasticity to buffer environmental changes. The environment is controlled under lab conditions, so loss of robustness due to genetic changes will almost never be detected. We have no good estimate — not even a ballpark! — of how many of these 95 “no obvious phenotype” cases actually have no phenotype, and beliefs widely vary.
    - tleonardi 7y ago
      I couldn't agree more. Many knock-out mice for protein coding genes don't have an "obvious phenotype", but they are conserved across species. When studied in detail they often reveal an effect, be it in the response to environmental stress or subtle behavioural alterations. I can't find an updated reference, but results published in 2012 by the Mouse Genetics Project and the International Mouse Phenotyping Consortium indicate that ~20% of coding gene knock-out lines examined don't have an "obvious phenotype" [1]. The most likely explanation for their conservation is that they do have a function and a phenotype: we just don't know what to look for. [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3463797/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3463797/

jballanc 7y ago

So, two quick points in response: 1. Yeah, I probably overstated the case. It's not as if it's impossible to move a gene somewhere else on a chromosome and have it function completely normally. As you point out, we do this all the time. That said... 2. Genetics is hard, and I'm of the school of thought that one reason it is so hard is because systems are more fragile in meaningful ways than we assume. That is to say, if you attempt a gene translocation 100 times and get 3 successes, most geneticists would be happy with that result and not question the 97 failures. And, indeed, it is entirely possible that the quality of the reagents, the stringency of the protocols, or any of an infinite number of other variables that have little or nothing to do with genetics could be the cause of those 97 failures...or, it could be that 97 times the translocation worked but the product wasn't viable. I know from experience that it is very hard to distinguish between these possibilities, and furthermore that in the publish-or-perish world of academia today there's not any motivation for navel gazing into the reason most genetic protocols are not more efficient. That said, I've seen hints in my past work that make me think location, gene dose, DNA secondary and tertiary structure, etc. are more consequential than the common wisdom would have you believe. (Oh, and for anyone still reading, I'll share the brain teaser that was my light-bulb moment for the extent of gene dose effect in prokarya: it takes 40 min to replicate the E. coli genome, but during exponential growth E. coli can double in number every 20 min. How?)

sjg007 7y ago

https://www.annualreviews.org/doi/abs/10.1146/annurev-genet-112414-055008 https://www.annualreviews.org/doi/abs/10.1146/annurev-genet-... and https://www.sciencedaily.com/releases/2010/10/101014144312.htm https://www.sciencedaily.com/releases/2010/10/101014144312.h...

dnautics 7y ago
The second one relates to a feature at the metaorganismal level; namely the evolution rate. Fun fact: human chromosome 1 is the merger of two simian chomosomes and has two centromeres

gravelc 7y ago

I'm going to disagree with go with 'sometimes yes and sometimes no'. Take a construct and transform it into a plant. Check the expression of the transgene from hundreds of different transformation events, and there'll be huge variation, from highly expressed to silenced, even though all inserted transgenes have the same promoter and terminator. Location matters. However, if the function of gene A is redundant, or it goes to a location when expression is not affected, or it's just not a very important gene, then it probably doesn't have any great phenotypic effect.

dnautics 7y ago
I would make a bet. Randomly generate 100 spots in the mouse genome, let's say maybe not in centromeric locations. Don't correct for being inside an ORF or a promoter. Pick 100 random translocations swapping these spots. 95 of them will be viable with no obvious phenotype.
- hyperbovine 7y ago
  In how many samples are you doing this? It’s all about effect size.
- klmr 7y ago
  I broadly agree with your point but careful when saying “no obvious phenotype”. As my former boss likes to say about work in C. elegans: “Worms can’t talk. They seem healthy but maybe they’re extremely unhappy.” Less flippantly, we know that biological systems in general have huge plasticity to buffer environmental changes. The environment is controlled under lab conditions, so loss of robustness due to genetic changes will almost never be detected. We have no good estimate — not even a ballpark! — of how many of these 95 “no obvious phenotype” cases actually have no phenotype, and beliefs widely vary.
  - tleonardi 7y ago
    I couldn't agree more. Many knock-out mice for protein coding genes don't have an "obvious phenotype", but they are conserved across species. When studied in detail they often reveal an effect, be it in the response to environmental stress or subtle behavioural alterations. I can't find an updated reference, but results published in 2012 by the Mouse Genetics Project and the International Mouse Phenotyping Consortium indicate that ~20% of coding gene knock-out lines examined don't have an "obvious phenotype" [1]. The most likely explanation for their conservation is that they do have a function and a phenotype: we just don't know what to look for. [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3463797/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3463797/