Determination of dominance or recessiveness

[nomadreid]

TL;DR

In pairs of alleles subject to the dominant/recessive distinction, can one determine a priori which one will be dominant?

Suppose you have two new alleles whose combination (gene) determine a certain trait, but the trait associated with only one of the alleles (i.e., those which can be compared as dominant/recessive). Of course the classic way, going back to Mendel, was to mix them and see what came out (statistically). But assuming one can sequence each allele before they are mixed (we are assuming that we don't know the outcome), can one, only by looking at the two sequences, determine which one will be determinant relative to the other? (As is certainly clear from my question, I am not a biologist, so the question is on a layman level.)

 

[BillTre]

This might work with some traits, but probably not all.

The important thing is what the trait is. Some, like the trait of carrying a specific sequence (a molecular trait) will be obviously dominant, since if you can detect the sequence in a heterozygote, it will be a dominant trait.

Other traits, of course, can be much more complex, like eye color.
Many genes affect eye color, with lots of subtle differences, in eye colors as phenotypes.
If the process generating a trait is not fully understood, or if the functioning of the gene with the sequence variants is not fully understood within the process of generating eye color, then it will be correspondingly more difficult to predict what a particular sequence difference will have on a phenotype. Seems unlikely to be successful to me for most fairly complex phenotypes.

This is assuming that the rest of the genome is fully understood and that sequence variants of other genes in the genome won't interact with the mutation differently. The sequence variance of other (possibly interacting) genes would add another level of prediction difficulty.

Making crosses or looking a family lineages vs. the distributing of the sequence variants would be easier.
From a experiemental point of view, you might have a cycle of hypothesizing what phenotype a sequence difference might result in, generating in some way an organism with the sequence variant, observe the result and determine if your prediction is actually predictive.

 

[Mitrick]

I think the best answer is No. Dominance is not easily predicted. For most genes we can't even predict whether the resulting protein will even be functional. Determining how two alleles will interact to produce the organism's phenotype is currently way beyond our capabilities. (Hi Bill)

 

[BillTre]

(Hi Bill)

Hi Rick.

 

[nomadreid]

Thanks very much, Bill and Rick.

 

[epenguin]

Your question is logical in the framework of the logic of causation. This is how these things may have to be presented to laypersons when you want to tell them something in a short time, because if you tell it in terms of the logic of discovery and what lead to the understanding that you are putting over, it is often a much longer story.

The logic of discovery has been the other way round from how you put it. The discovery of many mutations comes from the fact that a deleterious mutation, causing for example synthesis of a non-functional enzyme or other protein, is deleterious to the whole organism, and it is because of that it is first noticed. In the case that the protein is a human one (which is a large proportion of those studied ) it is a poorly functioning but not completely non-functioning protein that gets noticed, as it may be not totally incompatible with life but giving rise to a disease. Ones totally incompatible with life can easily not be noticed if heterozygotes are healthy and homozygotes never born. (In micro organisms you often conveniently do not have the diploid complication; the researcher can find haploid organisms mutant in an enzyme of a biosynthetic pathways, e.g. that of an amino acid by finding ones unable to grow unless supplied with that aminoacid.)

Starting around the middle 80s, researchers are in "protein engineering" able to synthesise genes with designedly altered nucleotide sequences, hence may make proteins with designed altered amino acid sequences at will. If you know the structure of the protein, and particularly that of the binding or catalytic site, you can try to predict the effect of changing the amino acid sequence.This prediction can be anything between a reasonable conjecture (certain changes will pretty obviously disrupt binding or catalysis) to a very advanced attempt at computer calculation of the folding of the protein. And so to some small extent what you suggest is actually done, but it is less as an instrument of prediction than a check up and way of improving our increasing prediction abilities.

I hope this is understandable, every sentence here would really require further explanations.That explains why it is easier to explain starting with the conclusions than how you get to them! Yours is also a good question it has made me realize questions that I do not know the answer to and I do not know if anyone does.

 

[nomadreid]

Thanks very much, epenguin. Very helpful.
I am well familiar with the fact that scientific and mathematical topics have two opposing ways of being introduced: (a) chronologically, as they were developed in history, and (b) logically. It is rare that (a) and (b) coincide.
I am also well familiar with the fact that there is a fine line between the amount of simplification needed to present an explanation for a lay audience and oversimplification to the point of falsehood. Your answer kept on the good side of that line, as far as I can tell.

 

[Eskcanta]

Suppose you have two new alleles whose combination (gene) determine a certain trait

That's not quite how alleles work, they don't both give "part" of a gene and together combine to make "a gene".

A gene that works and has a standard form, could be thought of like a "normal master copy" of that gene (not all genes in reality have one really common form that could be thought of as a "normal master copy", even if you just look at one species' form of the gene).

An allele is a complete copy of that gene, with or without any changes from what might be thought of as the normal form of that gene. Since you have two chromosomes, each with their own "complete" (including any errors, which can include large deletions) genetic code, you have two alleles of every gene, which might both be identical (homozygous) or not (heterozygous).

So really, each allele is by itself going to make its entire gene product, if it works at all (and that could actually be many different proteins https://en.wikipedia.org/wiki/Alternative_splicing). They won't work together to make 1 functional protein, like if you have one allele for just the first half of a gene and the other allele for just the second half, you get two versions of that protein both broken in different ways and not one functioning protein.

 

[Eskcanta]

Summary: In pairs of alleles subject to the dominant/recessive distinction, can one determine a priori which one will be dominant?

Kind of.

If one of the alleles has gained a new stop codon very early in the coding, you can predict it won't make the whole protein anymore. If the new product is easy to dispose of that allele will be effectively 'silenced', and the cell will probably make less of the functioning protein (using the allele that still works) than it had before. This can create a phenotype change, and it would be 'dominant', if you inherit the broken allele, you'll see the change, and in reading the allele's genetic sequence you can see the stop codon with just a little bit of training (especially if you have the right computer program helping you, and it's pretty easy to tell computers what to look for), and since it's close to the beginning of the protein you won't have to look long or hard to find it and you can safely predict it will cause a huge change in the protein, essentially turning that allele off.

All the other mutations in an allele I can think of would take more training or a better computer program to catch and explain. We're not very good at predicting outcomes to small changes yet.

 

[Ygggdrasil]

Kind of.

If one of the alleles has gained a new stop codon very early in the coding, you can predict it won't make the whole protein anymore. If the new product is easy to dispose of that allele will be effectively 'silenced', and the cell will probably make less of the functioning protein (using the allele that still works) than it had before. This can create a phenotype change, and it would be 'dominant', if you inherit the broken allele, you'll see the change, and in reading the allele's genetic sequence you can see the stop codon with just a little bit of training (especially if you have the right computer program helping you, and it's pretty easy to tell computers what to look for), and since it's close to the beginning of the protein you won't have to look long or hard to find it and you can safely predict it will cause a huge change in the protein, essentially turning that allele off.

All the other mutations in an allele I can think of would take more training or a better computer program to catch and explain. We're not very good at predicting outcomes to small changes yet.

The loss of function mutations that you describe are typically recessive. In many cases, the body can function well with only one copy of the gene. Sometimes this comes because of gene regulatory networks that can sense the body has a reduced amount of functional protein and increase transcription of the gene to compensate for lacking one functional copy. (In fact, some exciting research published earlier this year shows that the type of non-sense, truncation mutation you describe induces stronger genetic compensation that mutations/deletions that prevent transcription of the gene).

There, however, are cases where loss of function of one allele does lead to a phenotype. The situation you describe, where one functional copy of the gene is not sufficient for the body to work properly, does occur and is referred to as haploinsufficiency. Another way to get a dominant phenotype through a loss of function mutation is a dominant negative allele. A dominant negative allele is one where the mutant protein product interferes with the function of the normal protein product (often because the protein interacts with itself to form a complex). So, even if it is possible to identify a mutant allele as a loss of function allele, it is not always straightforward to know whether that allele acts in a dominant or recessive manner.

For a good discussion of how the biochemical effects of a mutation translate to its effects on phenotype, see https://en.wikipedia.org/wiki/Mutation#By_effect_on_function

 

[nomadreid]

Thanks very much, Eskcanta and Ygggdrasil. Very informative, leading me from my simplistic notions expressed in the question to deeper and more interesting facts. This includes your answers and the very useful links.