Proposal to evaluate genetic tools to evaluate Cattleya species

[geoffsharris]

New thread based on Is this jenmanii alba? discussion that developed out of is this plant Cattleya jenmanii or not?

The idea behind this thread is to think about how to use our collections to crowdsource plant material and financial support to pilot an evaluation of using modern genetic techniques to determine the genetics of various Cattleya species within our collections.

There are many questions that could be asked and answered. My personal goal at this point is to:

1) come up with a question of reasonable scope that it can be executed
2) get people to contribute as a community to the execution of the study
3) refine and demonstrate that this can be done on a small scale, but has the potential to be expanded much more widely to many other species and many more example plants
4) share with others with a mind toward growing the group interested in doing this
5) starting to map the genetics of famous clones whose genes are generally over represented in subsequent generations

I would love to hear from others to add to or expand on this basic idea, recognizing that this is intended to be a start small, show proof of principle and then expand idea rather than trying to boil the ocean all at once.

Can we pick a single or pair of species that are widely owned, have a number of clones with some genetic diversity out in the world and has some degree of question about a number of well known clones as to their genetic heritage.

Examples could be:

Cattleya labiata - there are two fairly separated regions in the wild and plants have demonstrated both different phenotypic traits and genetics. Can we get lot of examples of labiatas with as much parentage info as possible to see were the primary genetics of the plants in cultivation are from? In addition, I suspect that some of the labiatas in circulation have jenmannii or warnerii in their genetic history.

Cattleya walkeriana and lodigessii (or Cattleya walkeriana and nobilior) - there is clear gene flow between these populations. What are the actual genetics of the famous clones and current generations of plants in collections? This has to some degree been done here and is a good example of what we might consider doing https://www.scielo.br/pdf/asagr/v39n3/1807-8621-asagr-39-03-00315.pdf

Cattleya trianae - lots of oddities in the the old clones that have been used to make most of the modern crosses and plants occur over quite a wide geographic area with some significant phenotypic differences.

Cattleya mossiae and luedemanniana - while not a huge amount of gene flow between these, there is likely some and there are again very interesting differences in regional variants for the laraense vs. coastal lueddemannians

Cattleya jenmanii and labiata - very closely related plants both in growth, flower characteristics, bloom season, etc. is there an easy way to differentiate between these based on genetics?

Likely a thorny issue, but comparing tigrina and guttata in collections (and famous clones) vs. what is already known about genetic differences between these two species from the wild?

It would be great to get people to first suggest their interest in participating and what question do you think would be practical and interesting to tackle first.

I would suggest we leave out the idea of looking for purity of species if only because plants don't read books and neither do bees. There is some amount of gene flow between a lot of the cattleyas. We will rapidly find that most Cattleyas have their own version of small amounts of Neanderthal DNA either from currently existing (or extinct progenitor) species and there are a number of species are likely of hybrid origin to begin with. These are still wild species and merrily going about their business, but the species in itself has genes from elsewhere. To takle what is the reference species and how does my plant compare is headed in the boil the ocean direction.

I welcome your thoughts, encouragement, help and participation.

 

[terryros]

I am interested and would support financially. I would provide material from my modest collection of species and primary hybrids. I worry about knowing how much admixture is too much to qualify something as a species! While my genetics knowledge is better than average for human stuff, I don't think that extends into orchid world. I have no idea how precise genotyping of Cattleyas could be. I worry about the impact of plants that are mixaploids, even natural. What can we detect?

 

[geoffsharris]

Hi Terry,

The short answer is yes. I'm beginning to wonder if there is an even lower tech solution than fishing out genes for at least a first level screening. People are already using flow cytometry to just look at total nuclear DNA content to evaluate if plants are diploid, triploid and tetraploid, etc. I suspect that a decent number of the Cattleyas likely have enough variability of genome size in itself to tell the difference between somewhat distantly related species such as lodigessii and walkeriana for example. This may likely be true for leopoldii/tigrina and guttata as well based on their putative genetic history. There are services that are doing ploidy testing for $20 a test, so this might be an interesting first experiment, just simply looking at the ploidy of various clones and across several species to see if this could be used as a first level discriminator based on ploidy and genome size. Fairly low on the high tech scale, but potentially quite useful info and much easier to use straight away of knowing which clones are diploid, tetraploid or other combos that will likely render them lousy parents.

 

[terryros]

Replaces chromosome counting, which I am told is tedious and difficult to do well. I am not sure I have ever seen information on chromosome numbers for the various Cattleya alliance species. I have seen it for Phragmipediums. However, if my plant was analyzed and came back with more DNA content that a type species, it probably means EITHER that the species name is wrong OR it could be that it isn't standard diploid. We might not be able to tell the difference. I suspect that a number of our best species cultivars had increased policy?

 

[DrLeslieEe]

Thanks for starting this thread Geoff.

The above proposals might work.

However, I think for any good study of this nature, and not to repeat the previous work already done, we do need to know 3 things based on what I have read so far:

1. we need to know the ploidy status of the plant being utilized. As tough as the counting is, it must be done to determine 2n, 3n, or 4n status, even aneuploids. Without this knowledge, the extra chromosomes may warp the study.

2. Ideally the entire genetic code be map for the baseline species to use. Otherwise we have 90% or more similar genes overlap between two similar species and matching the loci or genetic markers to use is like shooting in the dark. It's like mapping out the human genome against the ape genome, and to determine what makes the human not an ape?

3. once step two is mapped out, then we need to map out one to three similar species for their genetic code. Then determine the differences between them all (aka which gene is unique only to that species). This is the only way to get the baseline to know the differences based on the comparisons. If species A has 3 sets that neither species B nor C have, then we an ID marker.

With this step we can then have that species genomes to compare with the 'doubtful' species. This will give more definitely pictures than if the markers were randomly selected as in the study by Lou Menezes on 'Kenny'.

Also with the complete genome, we can determine a species with DNA without flowers or plant before us, forensically.

Just thinking out loud.

 

[terryros]

I think you are right, Leslie, but it seems like a big reach right now, although I may have an inaccurate appreciation of the genetic tools available now for plants.

I looked once at the procedure for chromosome counting and the type of microscope it would require. I know it is not ridiculously hard, but was reluctant to do it myself at the time. When Geoff brought up flow cytometry I thought this might be a rough screen for a plant “other than a natural diploid” of the species. This ”other” plant could be 3N, 4N, xN.... It could also be another species, if various Cattleya species have distinctly different amounts of genetic material. Finally, it could be a hybrid. The number of chromosomes for the different Cattleya species must be known but I just haven’t seen it in print.

 

[geoffsharris]

Hi Leslie,

Thanks for contibuting!

My thoughts are perhaps along the lines of the lean startup methodology of developing the minimum viable product and then iterate and add complexity once we have proof of principle. I think we could have some short term goals in the pursuit of a grander goal. It will be a lot easier to develop the ability to ask and answer the question of is any given plant clearly or most likely of hybrid origin or is most likely the reference species. Demonstrating that an individual plant appears to have similar genetic variability to a wild reference set of plants and is therefore a pure species is a dramatically more complex task. This first approach could be a milestone on the way to asking and answering more complex questions.

To set the stage, all cattleyas (and formerly Brazilian laelias) have a baseline of 2n = 40. There are a couple of small rupiculous species that may be tetraploids as part of their evolution, but set those aside for now. All plants have chromosomal DNA and organelle based extra chromosomal plastid DNA. The plastid DNA is only transmitted by the pod parent much like mitochondrial DNA is maternally transmitted in humans. Several analyses have been done looking at highly conserved genes that are chromosomal (such as ITS) and organelle based (mat K, rpoB, ndhJ) that together are quite good in themselves at discriminating between species and have been used to derive phylogenic tree relationships and also highlight species that are very likely of hybrid origin.

1) Agreed that at baseline we need to look just simply at ploidy. Using flow cytometry today we should be able to both determine ploidy of individual plants with great accuracy, faster and cheaper than root tip squashes. Plants don't lose or gain single chromosomes (for the most part) so learning the ploidy of the plant and the total estimated genome size for several different species that are somewhat separated in evolutionary time scale would allow us to see just how much resolving power there is on genome size. Along with this, there are two ways to get tetraploid plants - first is that plants just simply duplicate their genome and have four copies of each chromosome vs. the standard two. This is what people are trying to do with colchicine treatments. The second is when a hybrid between two fairly different plants actually maintains copies of both parental genomes and now effectively has twice as many chromosome pairs and is more like 2n = 80 with a lot of redundancy between those chromosomes. This has happened a lot with various food crops humans use today. I don't think this is very prevalent in orchids, but as far as I know no body has gone looking at complex hybrids where this is more likely to occur than in a primary hybrid of closely related species.

2) I'd suggest we think about this from the other direction, if humans and apes are 98% the same, it makes more sense to evaluate the 2% not the 98% to decide which one is which. See above that there are identified genes that should have no or near no genetic variability within a species. By finding those differences we should be able to discriminate if a plant is jenmanii, warneri, gaskelliana or labiata just to use examples that are most difficult to tell apart from floral morphology. No doubt we could look for better and more discriminating SNPs if we sequence the whole genomes of multiple reference plants, but this is a massive task. I went to high school with the woman who started 23andme, and my concern is that this would be the orchid equivalent, requiring a lot more resources, participation, computing power and expertise to pull off today. Agreed that this won't tell us if 5 generations ago someone bred labiata to warneri and then backcrossed to one or the other so that only 3% of the plant is warneri genes and 97% labiata. I don't believe that this is that big of a problem with the plants that are owned by people today as these plants do us a favor by blooming at different times of year, such that someone would not accidentally mistake the two plants for the same species and then breed them. But this methodology should be plenty good to tell if mossiae 'Willowbrook', walisii 'Mt. Ito, trianae 'Mooreana' or the white jenamanii is actually jenmanii or are of hybrid origin. I personally think that focusing on a number of these famous awarded plants that have then been used for breeding or their descendents are probably the most interesting to tackle first.

I'd be super interested in seeing if we can come up with a few reference species well represented in our collections where we have a high degree of suspicion that we have both diploid and tetraploid plants (I have a bunch of lueddemanniana, trianae and gaskelliana that would candidates) and look at flow cytometry to assess ploidy and genome size to see if this would be a useful tool to discriminate between two or three not super closely related species. We could then go on to add more species over time. The outcomes could be that this technology isn't useful overall, is useful for telling ploidy, but genome size doesn't vary enough to be a discriminator or this could be a very useful tool for both ploidy assessment and species differentiation. See Genome size diversity in orchids: consequences and evolution

I'd also be interested in expanding out to sequencing once we have a good representation of diploid reference plants that are pretty likely to be the correctly identified as the species of interest. I'd also be happy to investigate a few labs that could do the work and scope out either doing ploidy and some amount of sequencing either serially or in parallel as makes sense from a cost and complexity standpoint. At the moment the goal would be to build on what we best believe are high probability reference plants of a chosen species and once we have that in hand we can decide if we can run that against some dubious plants.

As always, welcome everyone's thoughts.

 

[DrLeslieEe]

I do hear that we need to start small and build up. That is always the mandate of any small endeavors to find pragmatic achievable goals in baby steps.

The problem is we already have studies out there like the Menezes and Chase work that has set the stage to prove the theory is viable. They already have some markers that they have randomly screened to prove their point (multiple loci differences etc). We can technically already use these markers to prove the point of what it is 'not' rather than what it is. That is why a full genome study comparison between close species can wean out more similarities and differences to work with (rather than guessing from other people's work which has not been 100% accurate, leaving many unknown variables in the wind). We need a new standard that they don't have.

I agree it is a Herculean task and probably cost a bundle. But at least it won't repeat methods we already know exists.

What labs do full genome studies and what is the cost to do one baseline species?

 

[geoffsharris]

Found a nice paper to get the thinking going. A couple of years old, but that just means the cost has dropped even further. A guide to sequence your favorite plant genomes At this point sequencing a genome in itself isn't the cost driver. It's the human time to analyze the data and make coherence out of it. I suspect the combo of time, expertise and computing power is where we as a group will find challenges. That being said, we and others like us have the genetic material that is the interesting substrate. I think it could be super doable just as a 1.0 experiment to pick a species or two and just do ploidy and genome size. We are talking 10-15 samples of both likely diploid and polypoid examples of a single species. Once worked out, we would expand to other species. This body of work would be super helpful in just figuring out if a plant is diploid or poly and if it is likely the true species or not. Please see this link Plant DNA C-values Database | Royal Botanic Gardens, Kew and type in Cattleya for the genus. You will see that there has not been a lot of published data here, but that it points to the possibility of being able to figure out the diploid genome size and tell interlopers like walkeriana 'Pendentive' and 'Limerick' from straight walkeriana all just on genome size. We could do all 41 monofoliate and bifoliate Cattleyas with 5-7 examples for likely less than $5000 crowd sourced from our collections. This would be a hugely useful in that anyone in the future would for 15-20 dollars be able to figure out if they have a diploid or tetraploid plant and if the reference genome size is suspicious or not. Could even be made a requirement for future awards or allow for different awards for diploid vs. tetraploid plants. Food for thought.

 

[bethb]

If you're just hoping to look for phylogenetic relationships (who's related to whom, and how closely) and maybe do a little population genetics, you won't need to sequence the entire genome. You'd just need to select a few target genes, do a little rtPCR to amplify those genes, get the products sequenced, and then run some analyses (using software) to put together phylogenetic trees. The first step is definitely to narrow down the question you want to ask, but depending on what you want to know, you might be able to sidestep pricier methods (like whole-genome sequencing, which can be pretty expensive even using next-gen techniques) and still get the answer you're looking for.

 

[terryros]

For genetic analysis, flow cytometry or sequencing, is the source material growing root tips? How are the tips processed/stored/shipped to the genetics laboratory?

I would contribute hundreds of $ for an initial demonstration of what a lab can do with an orchid species (highly certain or suspect).

I have about 15 of the unifoliate species and with at least 5 of them, supposed tetraploids of the species. I know this is dwarfed by Geoff's and Leslie's collections so maybe my only contribution can be financial.

 

[geoffsharris]

I've spoken to a few of my other orchid friends, so I'm pretty sure we can figure out getting access to a great pool of plants. The first key will be to try and nail down either divisions preferably or mericlones of know or very high probability wild source material of a few species that are likely to have little introgression from other species. Walisii, gaskelliana, warneri come to mind for unifoliates, velutina, aclandiae, schilleriana for bifoliate examples. I'm starting to look into what we could do from the most complex of more or less sequencing genomes, to SNP profiling to just doing flow cytometry to look at ploidy and genome size. Once we have a better sense of what is realistic in scope we can think about how to execute it. The beauty of doing flow cytometry is that it is cheap and building up a database of genome sizes of various species would be pretty easy, you wouldn't need a particular reference lab and would be actionable by anyone who wants to test their plant to see how it compares to a know reference set of plants both in ploidy and genome size. SNP profiling is of intermediate difficulty and whole genome is an altogether different beast. Appreciate the support and enthusiasm

 

[DrLeslieEe]

Found a nice paper to get the thinking going. A couple of years old, but that just means the cost has dropped even further. A guide to sequence your favorite plant genomes At this point sequencing a genome in itself isn't the cost driver. It's the human time to analyze the data and make coherence out of it. I suspect the combo of time, expertise and computing power is where we as a group will find challenges. That being said, we and others like us have the genetic material that is the interesting substrate. I think it could be super doable just as a 1.0 experiment to pick a species or two and just do ploidy and genome size. We are talking 10-15 samples of both likely diploid and polypoid examples of a single species. Once worked out, we would expand to other species. This body of work would be super helpful in just figuring out if a plant is diploid or poly and if it is likely the true species or not. Please see this link Plant DNA C-values Database | Royal Botanic Gardens, Kew and type in Cattleya for the genus. You will see that there has not been a lot of published data here, but that it points to the possibility of being able to figure out the diploid genome size and tell interlopers like walkeriana 'Pendentive' and 'Limerick' from straight walkeriana all just on genome size. We could do all 41 monofoliate and bifoliate Cattleyas with 5-7 examples for likely less than $5000 crowd sourced from our collections. This would be a hugely useful in that anyone in the future would for 15-20 dollars be able to figure out if they have a diploid or tetraploid plant and if the reference genome size is suspicious or not. Could even be made a requirement for future awards or allow for different awards for diploid vs. tetraploid plants. Food for thought.View attachment 26792

That was a lot of reading and rereading until I got it. I'm afraid my genetics isn't up to date (more than 20 years ago when I was in university for sciences). However after digesting the info for a few days, I now have a grasp of what Geoff (and Beverly) is proposing.

I think initially we could definitely do ploidy (flow cytometry) and genome size. This can rule out polyploids easily. We just have to be careful that the genome size differences is not because of the ploidy issues. The difference in genome size can already separate walkerianas 'Kenny' and 'Limerick' from species status. Also to verify the validity of this test, 3 'wild' or jungle plants of the same species must be counted to make sure the SD of the results is acceptable.

Secondly, we can review the target genes that are currently known and create a database for them for each species that have been done already (Chase etc). I suspect there are many genes that are missing from these current studies that may be critical for the realistic delineation of species from suspected hybrids, particularly the ones that have been bred back for generations (and those undergoing introgression).

Thirdly, we can look at SNP's and loci targets that may need to be expanded and get those done with genome sequencing (instead of the entire genome), all species specific.

Lastly, we should really consider doing a full genome map of one cattleya species so that it can be a comparative model and the star of this project. People and institutions will fund something like this as it is easily understandable and is interesting to them. The type species of the unifoliates is labiata. Perhaps that can start the ball rolling. I am sure if I was at a function, I could talk people to fund this rather than a genome cytometry size comparison. It can be called 'The Labiata Project' lol.

Perhaps to get a true picture, multiple variables must be compared and checked with each other to come up with a reasonable logical deduction that can pass any scientific scrutiny.

Thoughts?

PS Terry, my understanding is that DNA are extract from growing tissue like root tips and meristems. But DNA is presently in all cells, so likely those were chosen for ease of segregation from other non-essential tissue.

 

[eds]

Been reading this with interest. I hope your proposal gets off the ground. My research at uni many years ago was sequencing and comparing mitochondrial DNA between Tanganyikan cichlids to determine their relationships.

I don't think this approach (using one or two sequenced sections of DNA between known primers) would be helpful in determining hybrids in orchids.

1. In hybrid introgression, some sections of DNA from each parent are mixed and, after a number of generations you may have very little of the 'contaminant species' DNA left. The chances of it being the section you are sequencing is small.
2. Depending on the rate of base pair change, you could have significant differences in these sections for different individuals of the same species, especially in different populations and it could give you differences due to natural change, not hybridisation.

Perhaps this paper may offer some suggestions better suited to detecting introgressive hybridisation?
https://www.nature.com/articles/hdy201168As this was 10 years ago perhaps these methods are now being used? (I'm a little out of date on this area now!)

 

[terryros]

Somewhere above Geoff noted that the unifoliate species all had chromosome counts around 40? In searching for this information I thought I saw some with lower counts, but am wondering if Geoff has a reference.

Pessimistically, if the unifoliate species have very similar (or even the same) base chromosome count and their differences are because of differences within the genome, I think that means that flow cytometry will only be able to tell us when a plant is more than diploid. It couldn't help with hybrid status.

Knowing that a plant was more than diploid with testing that was easier than manual chromosome counting would be an advance. However, we will need to do something more to deal with the hybrid issue if the different species have very similar chromosome counts.

 

[geoffsharris]

This is great. Please see Cassio van den Berg's thesis http://cassiovandenberg.yolasite.com/resources/thesis.pdf starting on page 51 for chromosome counts. The baseline chromosome count for the almost the entirety of the Laeliinae is 2n = 40. When you see number that are +/- keep in mind these are likely specific famous clones that are more than diploid or just the inherent error in doing root tip squash counts. As far as I am aware all the plants that are related to the unifoliate catts, bifoliate catts and plants related to Laelia/Cattleya purpurata are 2n = 40. For some of the plants that have now been recently lumped into Cattleya this may not be universally true, but these are infrequent exceptions.

The key part to understand is that while the normal number of chromosomes should be 40, the size of these chromosomes can vary tremendously from species to species. The number of genes themselves and the size of the translated portion of the genes will have only a small amount of variability across all of the Laeliinae as the genes needed have more of less the same function in all the plants. The proteins needed to make cellulose for example are universal, so genetic variability in such a gene or its absence just doesn't happen.

The difference is in the untranslated portions which can be hugely different. This is why you can have the same chromosome count but have big differences in overall genome size. Previously this has been termed junk DNA. It is anything but junk and is is likely one of the key ways that speciation has happened - accidental duplication of a gene and then it developing an altered function or expression pattern or untranslated portions that change gene expression patterns.

The question to be answered is are there enough differences between very closely related species (labiata, jenmanii, gaskelliana and warnerii come to mind) that they have enough difference in genome size to be quickly and easily differentiated just based on genome size. This is what nobody knows (or at least hasn't published). I'm quite confident that we will find big differences between somewhat disparate species such as warscewiczii and luedemanniana such that this will be completely effective to tell a hybrid between these from each true species. The question to be asked by measuring genome size is - Is there enough variability just in genome size to differentiate between all of the species or just some of the species of Cattleya? A baseline data set would allow anyone with interest and $20-50 to compare their plant to what is known for a species. For those trying to make breeding decisions this would be hugely helpful. Also would be helpful for determining if any number of famous outlier clones are outliers because they have a very rare allele or are outliers because they are primary or only modestly introgressed plants of hybrid origin.

The other thing we might find is that there are populations of plants that we as humans have designated as a single species might be better considered two closely related, but separate species - the two distributions of labiata come to mind and that some of the plants are in a tug of war of staying one species or becoming two - the coastal and larense types of lueddemanniana come to mind.

 

[terryros]

I get it! How do we identify a lab that is technically good AND willing to do it for us?

 

[geoffsharris]

Working on that now. There are quite a few labs that do this for other crop species and ornamental plants. Just a matter of figuring out who will deliver reliable data generation, be a good repository and the cost. This is where the start small and possible test a couple of labs might be the way to go to figure this out before making a grander commitment. The cost likely isn't going to be the issue, it's the hassle of sending samples in mass from different sources and coordinating that.

 

[geoffsharris]

Been reading this with interest. I hope your proposal gets off the ground. My research at uni many years ago was sequencing and comparing mitochondrial DNA between Tanganyikan cichlids to determine their relationships.

I don't think this approach (using one or two sequenced sections of DNA between known primers) would be helpful in determining hybrids in orchids.

1. In hybrid introgression, some sections of DNA from each parent are mixed and, after a number of generations you may have very little of the 'contaminant species' DNA left. The chances of it being the section you are sequencing is small.
2. Depending on the rate of base pair change, you could have significant differences in these sections for different individuals of the same species, especially in different populations and it could give you differences due to natural change, not hybridisation.

Perhaps this paper may offer some suggestions better suited to detecting introgressive hybridisation?
https://www.nature.com/articles/hdy201168As this was 10 years ago perhaps these methods are now being used? (I'm a little out of date on this area now!)

Thanks for the contribution Ed. I think we will likely need to set aside the question of species of hybrid origin or that are actually a complex mixture of genomes on their way to figuring out if it becomes one species over time, a species complex or stays two distinct species. This is at the core of what we as humans consider a "true" species. The plants don't care and are just doing their thing. My sense is that the real issue in our collections is down to what humans are doing with their toothpicks in their greenhouses in making crosses. This should be much easier to detect as the number of introgression generations is likely to be between 0-2. This is a fascinating element of orchid evolution none the less as there are a number of species that are very likely stable pretty homogeneous species today that at one point in time were fairly variable hybrid swarms.

 

[terryros]

Geoff, if the total DNA content by flow cytometry is going to be detectably different between species I can see that a tetraploid will have 2x the DNA of an established type species and be easy to identify and probably impossible to confuse with hybridization with any other species because none should have enough DNA to get that high. However, as we get to triploids and "mixaploids", I am thinking that we couldn't reliably distinguish these from hybrids with another species? They would be identified as "non-diploid" but we couldn't be sure what.