News for hackers. Stuff that matters.
Re:New Causes of Autism Found in Noncoding DNA by 781 on Tuesday May 28, 2019 @07:53PM (#20035612)
Stupid question from a noob: does gene location on a chromosome matter? If I swap gene A from chromosome 1 with gene B from chromosome 2, will this in general keep the cell viable and working the same way?
Re:New Causes of Autism Found in Noncoding DNA by jballanc on Tuesday May 28, 2019 @08:23PM (#20035767)
Gene location is extremely consequential, especially in Eukaryotes (everything that's not bacteria or bugs living near hydrothemal vents). A couple of things that are influenced by position:

* First, and most immediate, are promoter/suppressor sequences. These are the bits of non-coding DNA that regulate when a gene is turned on or off. If you move a gene away from its promoter, it will not turn on and off at the right times. These sequences are so closely tied to the proper functioning of a "gene" (typically a term that refers to the introns and exons of a coding sequence), that the operating definition of "gene" should probably be expanded to include them.

* Copy number. Chromosomes take a long time to copy, and if you had to copy the whole length of a chromosome from one end to the other you'd never get a chance to divide. So chromosomes contain multiple ORIs (origin of replication). Even so, it takes long enough to copy DNA that genes located near an ORI will have an effective gene dose higher than those located further away (i.e. it will be almost as if you have two copies of the genes close to an ORI and only one copy for those further away). Granted this effect is more pronounced in bacteria with circular chromosomes and a single ORI, where gene dose can differ by a factor of 4 (or even 8) depending on location, but the effect is still there.

* Sub-nuclear localization. This is an area that is still under active study (last I cared to look...which is a number of years ago now), but there does seem to be some order to the location of chromosomes within the nucleus during interphase (i.e. when the cell is not compacting chromosomes and lining them up to split the cell). It's likely that this localization is related to the distribution of various signaling pathways so that, if a gene is expecting to be "turned on" by a specific pathway, but it happens to become dislocated in the nucleus from where the elements of that pathway are doing the work of activating genes, then it may not respond the same way to extrinsic signaling.

* Finally, and most importantly: heterochromatin/epigenetics. In addition to regulation that occurs on a gene-by-gene basis, eukaryotes are also capable of large-scale regional regulation of their chromosomes. Essentially, a signal causes modification of the histones associated with a region of DNA and the entire region becomes "condensed" into heterochromatin. A gene in a heterochromatin region will not become activated even if all of the signals that would normally kick it into gear are going full blast. In other words, if you accidentally move a gene that should be active into a region of heterochromatin, you may as well have removed it from the cell entirely!

Re:New Causes of Autism Found in Noncoding DNA by dnautics on Tuesday May 28, 2019 @09:50PM (#20036182)
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".

Re:New Causes of Autism Found in Noncoding DNA by jballanc on Tuesday May 28, 2019 @10:39PM (#20036422)
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?)

Re:New Causes of Autism Found in Noncoding DNA by gravelc on Wednesday May 29, 2019 @12:44AM (#20036923)
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.

Re:New Causes of Autism Found in Noncoding DNA by dnautics on Wednesday May 29, 2019 @02:02AM (#20037211)
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.
Re:New Causes of Autism Found in Noncoding DNA by klmr on Wednesday May 29, 2019 @06:22AM (#20038498)
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.
Re:New Causes of Autism Found in Noncoding DNA by tleonardi on Wednesday May 29, 2019 @09:26AM (#20039860)
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/

Re:New Causes of Autism Found in Noncoding DNA by hyperbovine on Wednesday May 29, 2019 @03:20AM (#20037560)
In how many samples are you doing this? It’s all about effect size.
Re:New Causes of Autism Found in Noncoding DNA by dnautics on Wednesday May 29, 2019 @02:07AM (#20037232)
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
Re:New Causes of Autism Found in Noncoding DNA by ethanwillis on Tuesday May 28, 2019 @08:24PM (#20035769)
It does, genes are translated into protein products that affect cell operation. The relative concentration of these gene>protein products also impacts how a cell operates.

How efficiently genes are turned into proteins is determined by quite a few things, but one of them is definitely based on location. Upstream and downstream transcription factors will impact how many copies of a genes proteins are created.

Re:New Causes of Autism Found in Noncoding DNA by klmr on Wednesday May 29, 2019 @06:26AM (#20038513)
You could broaden your definition of gene to include proximal regulatory elements. If you take that unit and translocate it the answer is much less obvious.
Re:New Causes of Autism Found in Noncoding DNA by rcthompson on Tuesday May 28, 2019 @08:29PM (#20035794)
This is not a stupid question at all. In fact, it is the subject of much active research. The short and extremely oversimplified answer is that a given gene will produce approximately the same protein product no matter where it is in the genome, but the regulation of where and when and how much and under what conditions that protein is produced is highly dependent on the genomic context of the gene.
Re:New Causes of Autism Found in Noncoding DNA by polygonewild on Tuesday May 28, 2019 @08:06PM (#20035678)
Yes it does matter, see https://en.wikipedia.org/wiki/Chromosomal_translocation

Certain chromosomal translocations cause diseases.

Re:New Causes of Autism Found in Noncoding DNA by alwayslearning_ on Tuesday May 28, 2019 @08:31PM (#20035805)
The correct answer would be -- it depends. But in general, yes, the precise location of a gene on the chromosome matters.

The chromosomes are organized in an intricate three dimensional structure with different levels of hierarchical organization and interacting functionally. See https://en.wikipedia.org/wiki/Topologically_associating_doma... and https://en.wikipedia.org/wiki/Cis-regulatory_module

Re:New Causes of Autism Found in Noncoding DNA by tathougies on Tuesday May 28, 2019 @08:32PM (#20035811)
Yes, but it's dependent on a lot of things. Some chromosomal areas are more important than others. In humans and other mammals, it also matters from which parent the DNA came from (paternal DNA is different from maternal DNA, but a woman will 'rewrite' her father's chromosomes when passing it on to her child and a man his mother's, etc). Any variation in these things could cause developmental disabilities. However, typically things like inversions and translocations and such do not have huge phenotypic differences, whereas things like triploidy or uniparental disomy (inheriting both copies of a chromosome from one parent) have a more obvious manifestation.
Re:New Causes of Autism Found in Noncoding DNA by joshgel on Tuesday May 28, 2019 @09:26PM (#20036054)
Sorry this isn't accurate at all. We don't 'rewrite' any chromosomes. Sometimes during errors in replication changes occur, but these are random. They presumably happen approximately equally in both parents contribution to your DNA (though I don't know this to be 100% confirmed empirically).
Re:New Causes of Autism Found in Noncoding DNA by tathougies on Wednesday May 29, 2019 @02:25PM (#20043626)
Imprinting is a thing, and yes, the sex of the parent matters. A woman does not pass on a male imprinted version of her father's chromosomes. She passes on the female version. While methylation is not part of the 'genetic code', it is part of the DNA molecule, and it does impact genetic expression.

Anyway, in your pseudo-scientific world, how do you explain Angelman and Prader-Willi syndrome? Please... the world must know.

Re:New Causes of Autism Found in Noncoding DNA by mjg59 on Tuesday May 28, 2019 @09:56PM (#20036221)
Not in terms of primary sequence, but imprinting has an impact on expression and is parentally determined.
Re:New Causes of Autism Found in Noncoding DNA by sjg007 on Tuesday May 28, 2019 @08:05PM (#20035677)
Most likely yes. There's something called chromatin that winds up DNA. Also transcription factors and other factors play a role.