goldfish phylogeny

[fishes2catch]

Hi Everyone,

Does anyone know of a good goldfish phylogeny (genealogy)? There seems to be a bit of controversy in the literature. In J. Smart's book, 2001, there is a phylogeny (genealogy) that has dorsal fin loss and head growth arising independently. Where as in a recent paper by Komiyama in Gene, 2009, they find dorsal fin loss arising once. The later study was done using the mitochondrial genome. I'm interested in the independent fixation of the same traits in goldfish lineages. I'm just starting to dive into the literature. Does anyone have any suggestions? Having a good phylogeny will be essential for this.

Mark

[bekko]

Below is the genealogy chart from Matsubara, 1908. While Matsubara was an academic, not a life-long goldfish breeder, he did seem to have the cooperation of the Japanese breeders of his day.

I have read Komiyama and was not impressed. He seemed to be more interested in the process than the results. Anyone serious about goldfish could have come up with a more interesting experimental design.

A possible flaw in Komiyama conclusions is that the various varieties and individuals available to us today do not represent the results of rigorous divergent line breeding. There has been interbreeding of all the varieties many times over with subsequent selection to pull them back to a recognized variety.

Everyone breeding goldfish has noticed individuals which would be classified as a distant variety in the progeny of what appears to be "pure bred" parents. Do these odd-balls represent spontaneous reappearance of the same foundation mutation upon which the other variety is based? Or, do they represent a throw-back to inter-variety hybridization at some point in the past?

-steve


http://www.raingarden.us/geneology%20goldfish.jpg

[fishes2catch]

Hi Steve,

I agree with you about Komiyama's paper. This is a technique that they used on Japanese ornamental chickens and applied it to goldfish. It was more about the technique than the results.

I also agree that inbreeding between lines (or reticulated evolution) is a huge problem when trying to understand phylogenies. Are there any "pure bred" lines of goldfish available to study the phylogeny of goldfish?

Mark

[bekko]

Are there any "pure bred" lines of goldfish available to study the phylogeny of goldfish?


I suspect not - at least not for the major varieties. If there were, there is probably no documentation. But, then again, what is "pure bred" and what is a "line"? How many generations? With a thousand years of goldfish history there must have been nearly a thousand generations.

Orme (1979) describes the process of line breeding and developing two divergent lines for later out-crossing over ten generations. However, you see the process described in the same way for many domesticated animals (right down to how the table is formatted). So I suspect his recommendations are largely speculative and not developed from trial and error. Murphy's Law and human nature being what they are, I doubt that many tenth-generation line bred goldfish have ever existed. The consensus seems to be that you need to out-cross at about every fourth generation to recover from in-breeding depression.

When investigating the phylogeny of varieties, the first hurdle is defining "variety" as it relates to goldfish. Some Chinese sources say the number of varieties is over 300 while others say it is a about a dozen.

Certainly the degree of segregation is not a useful (or used)criteria for differentiating varieties. A single pairing can easily yield siblings which segregate into a half-dozen varieties (using the classification of Yuan, 2000). The fact is, every individual is different - so different that they can usually be recognized with the naked eye. The demarcation between a variety and just a group of individuals which share a common characteristic may be limited only by our ability to invent new names.

New varieties are typically new color combinations, recognizably different fins or minor differences in body conformation. Some new varieties are nothing more than marketing smoke as the industry tries to sell fish with characteristics which were traditionally discarded.

If you were a splitter and picked a new variety you may be able to find something with documented lineage over several decades.

-steve

[fishes2catch]

Hi Steve,

Maybe I'm thinking of goldfish in the wrong way. I consider them much like dogs. In dogs there are distinct varieties, e.g differences between great danes and chihuahuas. I see the same thing happening with goldfish. There is definitely distinct morphological differences between for example ranchus and shubunkins.

I'm a newbie as I have stated before but what I have read and talking with others, when crosses are done between ranchus you get ranchu offspring and not shubunkin (and the same is true for most "major varieties"). Is this not true? If it is true, I would consider this a variety or a line or what ever you want to call it. I know that there is a lot of variation on the "type" phenotype in the offspring and that some lines vary more then others but ranchu x ranchu = ranchu. The point is that (at least I think) there are "varieties" of goldfish that have distinct sets of genetics that distinguish them from one another.

I'm no splitter and I think that it is a bit silly that every new variation that comes out of a cross some are calling a new "variety". I do think that there is potential to create a phylogeny looking at the "major varieties" of goldfish, if my thinking is correct.

Mark

[Virginia ranchu]

Mark,

This is an interesting thread. I have spent quite a bit of time thinking about this and observing the inheritance patterns of goldfish.

I think it is important to recognize that goldfish are tetraploid, meaning they carry four copies of each chromosome rather than just two as in diploid animals like humans, dogs, cats, etc. This makes the number of possible gene (allele) combinations much higher, so "fixing" a strain is more difficult. Also, some of the types we breed for are heterozygous conditions, so it will always be impossible to develop a true breeding strain.

It is possible to breed some very striking fish by simply crossing types, but the development of a true strain requires generations of line breeding. Such hybrids will not breed true. Also, similar phenotypes in different varieties aren't always genetically similar. I once tried to develop calico TVR by crossing a calico lionhead with a metallic TVR. In the first generation, all the fry had dorsal fins.

Cheers,

Rob

[HNLim]

I once tried to develop calico TVR by crossing a calico lionhead with a metallic TVR. In the first generation, all the fry had dorsal fins.

So do they look like ranchu or oranda or short tail oranda?

[fishes2catch]

Hi Rob,

I think that it is very interesting that some of the phenotypes that we are concentrating on in goldfish are acting in a diploid manner. There is a huge body of research that is on going investigating what happens to genomes after a duplication. If you are interested, look up Axel Meyer's work. He and a fellow named Lou are comparing groups of genes in zebrafish, a close relative that is not tetraploid, to the same genes in goldfish to see what happens after a genome duplication. They are finding that many genes go silent (become pseudo genes).

Being tetraploid does make things a bit messy when trying to understand the genetics of traits. The lionhead x ranchu cross is interesting though. Essentially it is a complementation cross. What this tells me is that the genetics for dorsal fin loss is different in each line and that when combined they compliment and dorsal fins are restored.

Cheers,
Mark

[Virginia ranchu]

"So do they look like ranchu or oranda or short tail oranda?"

They had short fins like ranchu, but also a small dorsal. I kept some and I have crossed them to each other for two more generations. I am now getting some without dorsals. The other big challenge is recovering the melanin. By crossing to a red ranchu, the fry were red and white matts. Some fry are now keeping some black markings too.

Mark,

Thanks for the Axel Myer's reference. I will definitely take a look.

Rob

[bekko]

Rob,

I have often wondered whether lionhead and SVR are not more genetically similar than TVR and SVR. Being sort of a cult thing, TVR may have lived in relative isolation for quite a while. Your experience would indicate that this may be the case. When modern lionhead and SVR are crossed the dorsal is not restored.

But with that said, I raised a batch of metallic TVR crossed to metallic lionhead from a hand spawn Harris made. The intent was to strengthen the back and peduncle of the lionhead. The cull rate was very high - about 90% at 1/2 inch as I recall. There were a lot of saddle-backs, spikes and missing anals, but no complete dorsals.

-steve

[HNLim]

Many years ago, I bred TVR and there wasn't a single piece with dorsal or part of it.

[Cincy Ranchu]

Smart has had some interesting papers on this subject. Of all the ones published in the English magazine was the article about a recessive gene being isolated and then bred for several generation until it becomes dominant. The inference is that once dominant when outcrossed it remains dominant, eventhough previously it was recessive. In recent years we have seen this ( I think) this. Foster crossed a male Jikin to a Philadelphia veiltail and got a 100% of the fish to have the uniquely splt tails even though they were very long. This trait still remains in some of the veiltail lines to day. Circa 2005, a guy on this page in an effort to remove inbreeding issues with (crooked mouth syndrome)Bristols that had been isolated for many generations outcrossed to a American Calico Shubunkin. More than 90% had the familiar heart shaped tail in the first generation. Only the Matt fish ( calico Xcalico) yielded fish with pointy tails. Perhaps we can get Parker ot republish or at least send out these Scientific notes form Smart.

[bekko]

HNLim, I do not get spikes in my TVR offspring either. There are occasional spikes in the SVR and lionheads. But the out-cross of TVR to lionhead made a lot of spikes. It is sort of similar to Rob's experience, only without any complete dorsals.

Out-crossing provides the benefits of hybrid vigor and overcomes inbreeding depression, but you pay a high price for it.

-steve

[Virginia ranchu]

Gary,

I think the observation of a once "recessive" trait becoming "dominant" after inbreeding could be explained by the accumulation of multiple copies of that allele in a tetraploid species.

In a diploid, you will either have one copy, or two copies. In goldfish, you could have 1 or 2 or 3 or 4 copies. The inheritance of the trait could be related to the copy number.

Rob

[fishes2catch]

Hi All,

Copy number should not matter unless there is a dosage issue. This could be caused by differing genetic environments that the allele is placed. Genes/alleles are not islands and interact with other genes to produce a phenotype. An allele in one genetic environment could act as recessive but when crossed into another genetic environment could act as dominant.

Mark