Great Ocaspectations (2) Oca Genetics: This Much I know

OK - it's time to disperse the cloud of unknowing that obscures the summit of Mt Oca.

I'm going back to basics, just like former Prime Minister John Major did; I hope it works out better for me than it did for him. I make no apologies for using the technical terms for various processes - if I've got to learn them, so have you. "This Much I Know" could be rephrased more truthfully as "My Brain Hurts".

Brace yourselves.

Oca's genetic code is located on eight different chromosomes. In a "normal" diploid species they occur as pairs, which means 2 x 8 =16 chromosomes in total, see below:

In oca's case, however, the picture is more complicated: it is is a polyploid, with four sets of each chromosome pair, (2n =8x = 64). This means it is an ocatoploid, sorry, octoploid, with 64 chromosomes in total:

It is also an alloploid, formed when the different genomes of precursor species joined together. The latest evidence suggests that it may be an auto-alloploid. Resist the lure of a darkened room and bear with me for an explanation: this is gonna hurt me more than it hurts you.

What apparently happened is that one diploid parental species with the diploid chromosomal complement AA, doubled its chromosomes to form an autotetraploid AAAA (2n= 4x = 32, no chromosomes from any other species):

This then combined with two separate diploid species (BB and CC) or one allotetraploid (BBCC), itself formed by the uniting of two diploid species and nothing to do with the BBC, except perhaps in that they both contain a lot of repeats.

The result is AAAABBCC, the auto-alloploid Oxalis tuberosa we all know and love, with half the genome coming from species A, the other half from the B and C genomes, one quarter each.

The identity of these species is still not entirely clear, although likely suspects are Oxalis chicligastensis from northwestern Argentina and Bol/WT, an as-yet-unnamed Oxalis species from Bolivia.

Much of this information comes courtesy of Our Lady of the Ocas, Dr Eve Emshwiller, whose pioneering work on the genetics of Oxalis tuberosa is starting to unravel the mysteries of this crop. Her latest paper from 2009 gives much more detail on this and also has a lovely picture of a collection of oca tubers that are positively indecent in their diversity. Food porn I've heard of, but tuber porn (which this surely is), is a new and welcome addition to my life. Not such an inappropriate connection as it happens: oca has strong sexual associations in Quechua language and culture. The phrase oqa-tarpu, for instance, which literally means "planting ocas", doubles as a term for sexual intercourse. Another useful piece of information is that an oqaratu is a stupid, lazy or vacillating person. Feel free to spice up your domestic disputes about the gone-off milk, the missing car keys or the unpaid bills with this useful term of abuse.

In terms of character inheritance and plant breeding, where does this leave us? We know that oca is an obligate outcrosser (allogamous) with a system of self incompatibility based around three stylar morphs, short medium and long (tristyly). The genetics of this have been elucidated here. Basically, the inheritance of flower morphs was shown to be tetrasomic (involving four versions of the same chromosome at meiosis) and diallelic with two alleles of each of two genes: S,s and M,m, with S preventing the expression of the M gene - a process known as epsistasis. Various combinations of the above M and S genes and their alleles give rise to the three flower morphs. And it's all happening on the autotetraploid subsection of the genome.

Don't confuse epistasis with dominance. Dominance is caused by the dominant allele of a gene overriding a recessive one. Epistasis involves separate genes in which one has a cancelling effect "upstream" of the others - like trying to change channels on the telly when it isn't plugged in yet. Or (how's this for topicality?) head straight to Daughter of the Soil's post here for a practical demonstration.

So if all the flower morph activity (a big if) is located on the autotetraploid AAAA chromosome chunk, there must be a whole lot more going on over at the BBCC section of the genome. There will be plenty of genes located in these other ancestral genomes which may not be inherited tetrasomically. Which genes are where and which ones are inherited in which way are as yet, to me anyway, completely and utterly unclear. I feel the mists descending and the summit of Mt Oca disappearing from view. That or a migraine coming on.

It's probably a safe bet however, that oca is heterozygous for many traits and that these will segregate in subsequent generations in all sorts of complicated ways due to its mixed up genome. Oh joy. Mendel was lucky he chose peas. Nevertheless, I would imagine that the traits will fall into two categories, qualitative and quantitative.

Qualitative traits are those likely to be under the control of single genes or at least only a few genes: a binary on/off, present/absent, as shown in classic Mendelian inheritance (see Daughter of The Soil's blog for a thorough and lucid exposition of all of this).

Quantitative traits are ones which grade seamlessly from one to another along a continuum. They are under polygenic control, that's to say many genes, perhaps on different chromosomes, have an influence on any particular characteristic. These include, in some other tuber crops at least, traits such as yield, tuber shape and flesh quality.

I suspect, although I don't know, that tuber colour and plant form in general are maternally inherited - I should be able to put this to the test when I harvest the tubers in the autumn. Among the more obvious traits I can detect above ground are stem colour (red/green) underside of leaf (purplish/green):

And the absence or presence of axillary colouration. It would be fun to try and work out the mechanism of inheritance in these cases. If only I knew who their parents were.

Under the circumstances of my overwhelming ignorance, it is probably as well to follow the time-honoured route used by some potato and sweetpotato breeders - recurrent mass selection. It's a case of crossing varieties in every possible combination, then sowing and selecting their progeny, eliminating the feeble or substandard and keeping the best for the next breeding cycle. I could also do a little controlled crossing here and there and see what happens.

My main concern is that my seedlings have been derived from a handful of varieties, possibly less. Ideally you would start with a population of at least twenty unrelated individuals in the "crossing block". All is not lost however, as it should be possible to incorporate genes from new varieties by crossing them with the other ones to introduce more genetic variability - a process known as introgression.

There you go - that's close to the sum total of my knowledge of oca genetics at present: not a lot. That darkened room awaits me. What about you?