Apr 292017

It may only take a mammal 12 seconds to poop – but poo contains a treasure trove of information about the animal and its environment that can take years to unravel

All animals have to do it and according to recent work it doesn’t matter if you are a giant panda, dog, mouse or elephant – it takes about 12 seconds poop. While not the most pleasant sample to work with, there is a lot of information hidden in ‘poo DNA’. We can find out about the individual that produced it, the organisms (good and bad) in its gut and the plants and other animals in the environmental that it has eaten.

Every living thing contains DNA and leaves traces of that DNA in the environment. Poo DNA is a treasure trove of information that can be used to understand species, ecosystems and food-webs.

The animal that deposited the poo has left traces of its DNA. This allows us in determine which species of animal the poo came from, without needing to see the animal. We can even work out exactly which individual it was! This means we can track individual animals, or monitor the size of a population without disturbing them, just by collecting their poo. We can also find out about the health of the animal by the communities of organisms that live in animal’s gut, which can be beneficial, aiding in digestion, or parasitic / infectious just by sequencing the DNA found in poo.

BUT poo DNA also contains information on the diet of the animal – the plants and animals that have been eaten. This can tell us about the other species in the area and the food-webs within the system. For example understanding the different species of plants that an animal feeds on is critical to ensuring their habitat is effectively conserved, especially if they have highly specialized diets, such as giant pandas.

How it works

Once a sample is collected from the field the DNA is extracted. This is a mixture of DNA from all the different animals, plants and microorganisms that were in the poo sample.

To identify the individual species, we look at specific bits of the DNA that are good at telling species apart. These are known as DNA barcodes. Once we have sequenced these DNA barcodes we compare them to a reference set of DNA barcodes.  Reference DNA barcodes come from sequencing the DNA of known plants and animals that are usually held within herbaria and museums and have been identified by experts (taxonomists). The specimens within herbaria and museum are critical to making sure that we accurately identify which species are in poo DNA. It also means that we can go back later and do this for another animal in the same or different location and compare the results. Without this it would be almost impossible to track how an animal’s diet changes throughout the year, or understand the differences between locations.

When a DNA barcode from the poo DNA matches one in the reference DNA barcodes we know that the DNA of that species was in the poo sample and was probably eaten by the animal that deposited the poo. This tells us that not only is that species in the area, but that it is a potentially important food source for the animal we are studying.

The Royal Botanic Garden Edinburgh is using this technology to investigate the diet of giant pandas. Sound simple? Well it actually might be more complex than you think. Check out our Botanics Story on Using DNA to understand giant panda diet.

Apr 122017

Once we realised that most of our plate of Schistidium ITS2 amplifications had been successful, it was an easy decision to process them all for DNA sequencing. If a higher proportion had failed, we would have had to “cherry pick”, selecting and transferring the successful reactions into new tubes. Every sequencing reaction has a cost, and so deliberately sending a lot of failed reactions through the process, knowing that they won’t generate DNA sequences, is worth avoiding. However, transferring the reactions into new tubes is a meticulous job. In a plate of reactions, there are 96 wells arranged in 12 columns (1-12) and eight rows (A-H).  The first sample is at 1A, then 1B, 1C, 1D, 1E, 1F, 1G, 1H, 2A… all the way to 12H. Imagine if the first failed sample is at 1B and the second at 1G – then we’d transfer 1A to 1A, 1C to 1B, 1D to 1C, 1E to 1D, 1F to 1E, 1H to 1F, 2A to 1G, etc… it’s easy to see that any interruptions can be fatal (this sort of task is why people in the lab sometimes have a “Do Not Disturb” sign stuck on the back of their lab coats). Thus, with only a few failures, we’re better keeping any liquid transfers simple (and manageable using a multichannel pipette).

A 96-well plate on ice

After electrophoresis and visualisation of the Schistidium ITS PCR products under blue, or U.V., light, the next step is to clean up these PCR products. This is where we remove unincorporated primers and dNTPs from the reactions. In our PCRs, we added all the ingredients in excess, so that there was more of everything than the reaction needed. This might seem wasteful, but compared to the costs of the time and plastics that would be required to carefully optimise each reaction, throwing lots in and getting quicker results really isn’t too profligate.

When a PCR is successful, the primer (an oligonucleotide, or short single stranded DNA molecule) binds to the end of the region that is getting copied, and a new strand of DNA is synthesised from the end of the primer, incorporating the primer into the new strand. Thus, primers get used up. The building blocks for DNA synthesis, the nucleotides (As, Ts, Gs and Cs), are the dNPTs (deoxynucleoside triphosphates) that we added to the PCR reaction, again in excess. So even when we’ve built a lot of copies, we still have some primers and dNTPS that weren’t used up.

This didn’t matter when it came to looking at the results on the gel, but does become important in our next reaction, the sequencing PCR. In the sequencing PCR, we add just a single primer, and we add a very precise blend of nucleotides. So we don’t want to carry over primers and dNPTs from the previous reaction. There are several ways of removing these.

The cheapest is through ethanol precipitation of the synthesised DNA, where the unincorporated primers and dNTPs stay in solution and are thrown away. This is less easy to scale up to plates, and the moment where you turn it all upside down and toss the liquid out is rather worrying – the cost of plates of subsequent sequencing failure if the DNA pellets were lost is huge, and the pellets tend not to be visible.

For several years we used a column-based approach, a bit like our DNA extraction method, where the PCR product is bound to a membrane, and the unincorporated primers and dNTPs are flushed off it, before the PCR product is eluted back into solution.

However, the method we now use most regularly involves a combination of two enzymes, Eco1, and Shrimp Alkaline Phosphatase (SAP), which work elegantly in combination, and which involves methodology that is easily scalable for working with plates and multichannel pipettes. The Eco1 enzyme digests single-stranded DNA, so it cuts the primers up into individual nucleotides. The SAP enzyme dephosphorylates the nucleotides (the dNTPs) so that they are no longer functional. Thus after using these enzymes, although nothing has been physically removed from the tubes, the unwanted reagents have been rendered unusable. For convenience, we buy in a commercial combination of the two enzymes called ExoSAP-IT™, and add a small amount of that to our reactions (in this case, around 1 μl ExoSAP-IT to about 15 μl of PCR product).

The enzymes work best at 37ºC, and so the reactions were put on a heating block for 15 minutes. The enzymes are killed at a higher temperature, and so the next step was to heat the reactions to 80ºC for 15 minutes, to make sure that no viable enzymes are left to interfere with the sequencing reaction. After that, the plate of cleaned Schistidium ITS2 amplicons was left in the fridge until it was needed for the sequencing reactions.


Links to reports on Moss diversity in an artificial landscape, an EU Synthesys Access project with Dr Wolfgang Hofbauer at RBGE:

Apr 112017


Gel loading in Lab 33

Once the polymerase chain reaction is over, it’s time to Run The Gel; this is make-or-break time, when we find out if our PCR amplification has actually worked.

The first step is to prepare and pour the gel. The gel is a 1-2% mix of agarose in a salt-containing buffer solution (for us, this is usually 1x TBE); the agarose is dissolved into the buffer using heat, in a microwave, and a bit of stain is added. We use about 5 μl of SYBRsafe for a 100 ml gel. The SYBRsafe binds to DNA and fluoresces under blue light, so we can see if we have produced amplified DNA.

After the agarose gel has cooled down to a comfortable temperature (so we don’t risk warping our ridiculously expensive plastic gel set-up with it!) we pour it onto a plate, and add a “comb” which sits in the gel and around which it will set, forming lines of small oblong holes in the gel’s upper surface into which we will later pipette our PCR product.

Setting up the loading buffer and PCR product in Lab 31

While waiting for the gel to set, we mix a small amount of the Schistidium ITS PCR product (3-5 μl) with a little loading dye (1-2 μl). The loading dye serves two purposes: it contains glycerol, which is rather dense and helps the liquid drop into the holes (wells) in our gel, and it also contains a colour (in this case, bromophenol blue), which means that we can see what we’re doing – dropping a colourless liquid through another colourless liquid into a colourless hole would otherwise be rather more an act of faith!

Using a manual multichannel to load the PCR product/loading solution

When the gel has set hard, we lift off the comb, taking care not to rip the gel (which is rather more brittle than gelatine-based jellies), and put the gel into a gel tank, covered with a thin layer of the TBE buffer that was used to make the gel. Then it’s time to load. Down one side, we put 3.5 μl of a “ladder”, a bought-in product containing DNA fragments of known lengths and quantities, which will act as a standard for the gel and let us estimate the sizes of any PCR products that show up. Adding the ladder, in which we know there is DNA, is also like a “positive control” that we’ve made our gel properly – if nothing at all shows up in the image of the gel, then it’s not proof that our PCR reaction failed, but that there was something wrong with our gel.

With the gel set-up that we are using for our plate of 96 Schistidium amplifications, we can load 8 samples at a time using one of the multichannel pipettes. For this, instead of our usual electric multichannels, we use a manual one so that we have more control over the rate of dispensation, and so we can stop if some of the liquid is not going in the right place: it’s quite a precarious operation!

Running our gel

Because the sugar-phosphate exerior of a DNA molecule has a negative electric charge, DNA molecules migrate through an electric field towards the positive electrode, and so if we apply a current to the gel, we can get the DNA to move through it. The filtering effect of the set agarose in the gel means that the DNA molecules migrate at a rate proportional to their size, with short molecules moving quickly, and longer ones more slowly. Generally, 30-45 minutes at 80 volts is enough to see if our PCR is successful.

After electophoresis, the gel is placed on the transilluminator

After the gel electrophoresis, the gel is carefully lifted out of the buffer, off the gel tank, and placed on a plate in a light box. (At this point with larger thinner gels like the one we ran for the Schistidium PCR products, it’s quite easy to break the gel.) Because the shape of the gel and of the light box are different, these gels need to be cut in two. Luckily, the halves fit nicely side by side in a single image.

Our gel beneath the filter

An orange filter is added, and when it’s turned on, the transilluminator shines with blue light that will make any DNA stained with SYBRsafe fluorescent, at which point a digital image is taken of the gel.


The first image of our Schistidium PCR products

To see what’s on the gel, we use an automated setting, where the computer choses the best exposure for the image. However, a bit of subsequent cropping and fiddling with the image itself can make for a far better gel picture. And the really good news for us is that, looking at our gel, it is clear that most of our Schistidium DNA extractions were successfully amplified for the ITS2 region in our PCR reaction!

The wells can be seen at the top of each row, as darker shadows; for the purposes of this image, the DNA has migrated towards the bottom of the screen. For only about seven of the 96 reactions is there no visible band of PCR product on the gel, and in the vast majority of cases a clear bright single band is present. (The fuzzier band that is sometimes visible a little further down than the ITS product, especially for the second last sample on this gel, is primer-dimer, an artifact caused during PCR, and not the amplified product we want; it’s far far shorter than the region we’re working with, as can be seen by comparing it with the size of our 1 kilobase DNA ladder).

Cropped and adjusted gel image picture, with a ladder on the right hand side of every third row

Links to reports on Moss diversity in an artificial landscape, an EU Synthesys Access project with Dr Wolfgang Hofbauer at RBGE:

Apr 062017

PCR set-up with defrosting reagents in Lab 32

After we extracted a plate’s worth (12 columns by 8 rows, or 96 samples) of Schistidium DNA, the next step in our process is to copy a preselected part of that DNA, using the Polymerase Chain Reaction (PCR). For this study with Wolfgang, we are copying a region of nuclear DNA known as the Internal Transcribed Spacer 2 (ITS2). We set up the reactions in a laminar flow hood, which blows clean air down onto the work surface, keeping everything as clean as possible. Most of the reagents are kept in a -20ºC freezer between uses, so these get set out to defrost before use (time for a coffee break!). The DNA polymerase enzyme, taq, on the other hand, is stored in glycerol so doesn’t freeze at -20ºC; while we’re using it, we keep it on ice so it stays cold.

The reaction components are added into a single “Master-Mix” tube, which includes water, buffer, Magnesium, enhancing additives, short oligonucleotides (primers), and the taq enzyme. Small (19 μl) aliquots of this master mix are then added to a new 96-well plate, and 1 μl of DNA from the extraction plate is transferred across into each of the 96 reactions.

Using the multichannel pipette to transfer DNA

Once the DNA has been added, the plate is sealed (this time with a clear plastic film that sticks firmly in place when heated), briefly spun in a centrifuge to make sure all the reagents are mixed together at the bottom of the 96 plastic wells, and transferred into one of our Thermocyclers, or PCR machines.

PCR plate in Lab 32 thermocycler

The plate sits in a metal block which rapidly “cycles” up and down in temperature, following a predetermined programme. For the gene region that we are copying here, the programme first heats the block to 95ºC for 4 minutes, then starts a cycle of 94ºC for 1 minute, 55ºC for 1 minute, and 72ºC for 45 seconds, repeated 30 times. When the block is at 94ºC, the double-stranded DNA is pulled apart into single strands; when it’s cooled to 55ºC the primers stick on and the taq initiates making copies of the ITS2 region, and when it is heated up again to 72ºC, the copies get completed, before they’re pulled apart again when the block heats up to 94ºC and the cycle starts again…

The thermocycler screen showing the ITS2 PCR programme

A couple of hours later, the reaction is complete, and at this point we HOPE that all the 96 wells in our plate contain millions of copies of the ITS2 DNA molecule from the Schistidium extraction that was added in each case. However, in order to see if any of it has actually worked, we need to stain and visualise the DNA, and for this, we have to run a gel, a process that will be the subject of the next installment.



Links to reports on Moss diversity in an artificial landscape, an EU Synthesys Access project with Dr Wolfgang Hofbauer at RBGE:

Apr 052017

Wolfgang hard at work in the Cryptogam Workroom at RBGE

Just over a week into our current Synthesys-funded Schistidium project, and Wolfgang has picked through piles of packets of mosses, selecting the 96 that we would most like to get DNA sequences for, and putting tiny pieces of them into plastic tubes for DNA extraction.

Ready to start our DNA extractions in Lab 30

Down in the basement, we have a storeroom with racks of lab coats, so we ventured down and found one that was a good fit for Wolfgang, to keep him from spilling chemicals down his regular clothes. We also got him some gloves; at RBGE, instead of latex, to which some people are allergic, we use nitrile. These serve two purposes – one, clearly, to protect our hands from any hazardous materials we might use, but in many cases in the molecular lab, the gloves are actually to protect our samples from us, so that we do not get our DNA or enzymes into the tubes we’re working with.

Our 96 moss samples ready for grinding

The first step of the extraction is the mechanical disruption of the plant tissue. Back to Biology 101, and a major difference between animal and plant cells is that plant cells have a tough cellulose-based cell wall around the outside of the cell membrane. To break through this cell wall, we add small tungsten beads to our tubes, and pop them into a TissueLysis machine which vibrates the tubes rapidly back and forth until the plants are rendered to a fine powder.

About half a millilitre (420 μl) of a salty, soapy buffer solution is added to this powder. Because we’re using tubes that are in strips of eight, we can use a multichannel pipette to dispense eight aliquots of the buffer at a time, reducing the amount of handling time. This first buffer helps break open the plant cell membranes, releasing the Schistidium DNA into solution. We put the tubes on a heated shaking block, leaving the samples for 1.5 hours at 65°C.

Schistidium DNA extractions with added binding buffer

After the lysis step, the strips are centrifuged to remove bits of plant debris, 220 μl of the liquid Schistidium extraction is pipetted into a new deep 96-well plate, and 440 μl of a binding buffer is added, before 600 μl of the solution is transferred to yet another plate… This next plate contains a membrane to which the Schistidium DNA binds. While the DNA is stuck to the membrane, we pass four wash buffers over it, to remove compounds that might have co-extracted with the DNA, but that might inhibit some of our downstream reactions. These wash buffers contain quite high proportions of alcohol, so they don’t remove the DNA from the membrane (because DNA is not soluble in alcohol it doesn’t come out of precipitation).  The final step in the process is to remove the DNA from the membrane in an elution step, using 100 μl of a water-based elution buffer per sample, in which the DNA can dissolve.

Our plate of Schistidium DNA, sitting in the Lab 32 fridge

Once the 96 Schistidium DNA extractions are finished, the plate is sealed (with a thin sticky metal film), and left in the fridge (in this case, as we are going to use the DNA the next day), or in the freezer, until we are ready to move on to the next step.


Links to reports on Moss diversity in an artificial landscape, an EU Synthesys Access project with Dr Wolfgang Hofbauer at RBGE:


Mar 292017

Some Schistidium collections from the RBGE Herbarium

Monday 27th March was the start of a month-long visit to RBGE by the Fraunhofer Institute for Building Physics‘s Dr Wolfgang Hofbauer, funded by the EU Synthesys Access programme. This funding enables researchers from other institutes to get their hands on the natural history collections that they need to see and understand, but it is equally vital for collections-based institutes like ourselves, as it promotes the use and curation of some of the material that we conserve.

DNA barcoding publication resulting from Wolfgang’s previous visit to RBGE

Wolfgang first visited us at RBGE as part of an earlier Synthesys programme, in 2014, which initiated a very useful collaborative project looking at the growth of species from the moss genus Schistidium on the built environment. We used DNA sequence data to try to identify some of these mosses, because the harsh environment in which they grow means that the plants are often malformed or underdeveloped, and difficult to identify using morphology.

Wolfgang has come back to RBGE in order to continue this work, in part by adding to our “Reference Library” of DNA sequences from different Schistidium species, but also to look at ways of developing our ability to grow some of these mosses where and how we want them.

We hope that by the end of this visit, we will be closer to answering the five following questions:

a) What is the taxonomy of Schistidium diversity on modern building surfaces?
b) Can we show geographic patterns of morphological and genetic variation in Schistidium on modern buildings in different European regions?
c) Are certain Schistidium taxa confined to special ecological situations (material, exposure, etc.) and can we use this for management of moss growth on buildings?
d) Which taxa of Schistidium are the best candidates for moss gardening, and is there potential for specially developed masonry/techniques to facilitate their growth?
e) Does improving the taxonomy of Schistidium on building surfaces allow us to find specific natural antagonists that could be used for its biocontrol?

Wolfgang hard at work in the Cryptogam Workroom at RBGE

The outcomes that we would like to see from this work are:

1) A complete baseline DNA-barcode library of Schistidium species as a tool for identification of sterile or morphologically atypical material.
2) Increased insight into the ecology and taxonomy of Schistidium species that grow on modern building structures.
3) A common project with the theme of deliberate growth of suitable cryptogams on building surfaces, as a collaboration between Science and Horticulture Divisions here at RGBE, and the Fraunhofer Institute for Building Physics, where Wolfgang is based.
4) Preliminary information on potentially specific biocontrol of unwanted growth on building surfaces, by identification of the moss lineages involved.
5) Development of an accessioned living collection of Schistidium species that have been identified using DNA barcoding and cultivated at RBGE, to be used for moss cultivation experiments and for public display.
6) Working with RBGE’s bryologists and horticulturalists, the development of a living display of moss colonisation (a kind of “living poster”) that can be used for outreach activities.
7) Published research, both on the DNA barcoding of Schistidium and the diversity of the genus on building surfaces, and also on bryophyte cultivation methods.


Links to reports on Moss diversity in an artificial landscape, an EU Synthesys Access project with Dr Wolfgang Hofbauer at RBGE:


Dec 022016
Some of the herbarium collections of Marchantia held in the RBGE herbarium

Some of the herbarium collections of Marchantia held in the RBGE herbarium

Many new species are already included in natural history collections around the world, it’s just that nobody has yet got around to examining the material, recognising that it represents something novel, and publishing a name for it. Sometimes these new species are filed under the epithet of a similar named species, sometimes they’re just filed under the genus name with other collections that have not been identified to species, and sometimes they have been annotated to recognise that they’re probably distinct from all the species that have already been described, e.g., as “sp. nov.

David Long has made a huge number of plant collections from around the world in his 40-plus year botanical career, with many of these collections not yet fully examined. Some of this material is being mined for DNA sequencing projects at RBGE, and for some of our key plant groups, as well as sequencing well-identified material, we are also sequencing plants that have not been assigned to species. Molecular lab work is fast compared to close morphological studies of multiple plant specimens; this can therefore speed up the processes of traditional taxonomy, by allowing it to focus on things that are obviously distinct.

One lineage that David Long is particularly involved with, and that remains one of our key plant groups, is the complex thalloid liverworts. Some of our sequencing work has involved Marchantia, which made Xiang et al.‘s recent description of a new species in the genus, Marchantia longii, particularly interesting. In the last few days, the DNA sequences that were included in the paper were made publicly available on the NCBI site, GenBank. One of the regions that was sequenced by Xiang et al., the plastid-encoded RuBisCo Large subunit gene rbcL, was also included in our study, and so I was able to put the two data sets together, and see how the new species fits into our phylogenies.

The results are interesting: When Xiang et al. named M. longii, they did so in part because the area that the plant came from, in northwestern Yunnan, is one in which David has been very active. In fact, at RBGE we had already generated DNA sequence data from nine accessions of Marchantia that David had collected there. I was delighted to find that two of these accessions (collections Long 36155 and Long 34642), which had been filed in our collections without a specific epithet, are an exact genetic match to Marchantia longii. It seems that David really does have an affinity for the plant, having gone out and found some even before it was named for him!


Long’s Marchantia

A rapid phylogeny of Marchantia, from the RBGE collections. II. Illuminating our sampling

A rapid phylogeny of Marchantia, from the RBGE collections. I. Sampling

Sep 142016
Sphaerocarpos texanus and S. michelii, from the British Bryological Society Field Guide (see references for link)

Sphaerocarpos texanus and S. michelii, from the British Bryological Society Field Guide (see references for link)

In conjunction with Dr Daniela Schill’s monographic work on Sphaerocarpos, we’ve been building a molecular phylogeny for the genus. We have attempted to extract DNA from 66 accessions, including three S. cristatus, all from California, seven S. donnellii from the US, five S. drewei from California, two S. hians, 13 S. michelii from France, Great Britain, Italy, Malta and Portugal, two S. muccilloi, five S. stipitatus from Nepal, Portugal and South Africa, and 25 S. texanus from Belguim, France, Great Britain, Italy, Portugal, Turkey, California and Illinois. We have also included some as yet unidentified material, including an accession from Chile.

Because much of our work at RBGE focuses on plant DNA barcoding and the protocols are established and frequently successful, we have chosen to use sequence data from some of these barcoding regions for this project. However, the liverwort matK primer sets were not very successful in Sphaerocarpos, with a very limited number of good quality sequences generated. The nuclear ITS2 region had its own issues, with many of the sequence reads being difficult to interpret due to overlapping peaks. In the end we focused on the three most successful plastid loci, the rbcL and rpoC1 barcoding amplicons, and a region that encompasses part of the psbA gene and the psbA-trnH intergenic spacer.

Unfortunately, we have not been able to amplify DNA from all the samples we extracted, with failures particularly for some of the older specimens. One of the species we attempted to sequence, Sphaerocarpos muccilloi, has not worked for any of the gene regions that we have been using, while another species, Sphaerocarpos hians, has so far only amplified for a single region, rbcL.

Sample phylogenetic tree for Sphaerocarpos, based on rbcL sequence data

Sample phylogenetic tree for Sphaerocarpos, based on rbcL sequence data

Although each species seems to be genetically distinguishable from the other species sampled, two of the most widespread species, S. michelii and S. texanus, are resolving as para- or polyphyletic. The phylogenetic tree contains three distinct groups of S. michelii accessions, and two distinct groups of S. texanus, one from Europe and the other from California. An Illinois accession that has been published as S. texanus resolves here with S. donnellii. The Illinois material lacks spores and is thus difficult to identify morphologically, but is outwith the Southeastern Coastal Plain area where S. donnellii is thought to occur.

The next steps in this study involve a second pass through the DNA extractions, to see if using other PCR additives will help increase the sequence success rate, then combining the sequence data from the three sequenced loci into a single matrix, to produce a more robust and supported phylogeny. Description of new species, where required, will fall under Daniela’s remit, in line with the comprehensive taxonomic revision that she has carried out.


Links & References:

Sphaerocarpos, preview to a monograph

BFNA | Family List | BFNA Vol. 3 | Sphaerocarpaceae

BBS Field Guide Sphaerocarpos michelii / texanus

Sep 092016
Sphaeropcaros texanus photographed by David Long (Long 33162)

European material of Sphaerocarpos texanus, photographed by David Long (Long 33162)

The Sphaerocarpales (or “Bottle Liverworts”) form a very distinct group in the complex thalloid liverworts, with ca. 30 species in five genera: originally the group just included Geothallus (monospecific), Sphaerocarpos (8-9 species) and Riella (ca. 20 species), with two more monospecific genera, Austroriella and Monocarpus, added within the last few years. All five genera have very unusual, and highly reduced, thallus morphologies. With the exception of Monocarpus, they also all enclose their sex organs (or gametangia – the antheridia and archegonia) in inflated flask-shaped bottles (as can be seen in the accompanying photograph). This feature sets them apart from all other liverworts. All of them are adapted to extreme habitats, including arable fields, hot arid regions, seasonal lakes and pools, and salt pans.

A worldwide revision of the second largest genus of the group, Sphaerocarpos, is over 100 years old (Haynes 1910); other revisional work focuses on individual geographic areas, including South Africa (Proskauer 1955), North America (Haynes & Howe 1923, Frye & Clark 1937, Schuster 1992, Timme 2003), California (Howe 1899), Europe (Reimers 1936, Müller 1954), and France (Douin 1907). No revisions have been made for large areas including Australia, Asia and South America, and most of the work predates any DNA-based concepts of plant identification or species relationships. Bringing the taxonomy of Sphaerocarpos into the 21st century, Dr Daniela Schill spent 18 months (2007-2009) at RBGE on a Sibbald Trust-funded project to compile a world-wide taxonomic revision of the genus. Two field expeditions fed into the project, with Dr David Long collecting European species in Portugal in April 2007, and Daniela collecting North American species in California in March 2008 (funded by the Peter Davis Expedition Fund).

Spore SEMs of Sphaerocarpus drewiae, taken by Daniela Schill

Spore tetrads of Sphaerocarpos drewiae, SEMs taken by, and plate prepared by, Daniela Schill

Daniela’s work is based on morphological and anatomical characters, including spore characters that she observed using Scanning Electron Microscopy (SEM). Her aim has been to produce identification keys to the species, species descriptions, species lists, synonyms, botanical drawings, distribution maps, and ecological, nomenclatural and taxonomical notes. Although the study is not yet published, much of it, including SEM plates for spores from the ca. 9 different species (as seen on the right), is complete.

In parallel, RBGE staff have also been sequencing multiple accessions of all available Sphaerocarpos species, producing data that has helped inform some of Daniela’s taxonomic decisions, and that also allow us to generate a stand-alone phylogeny for the genus.

This research will lead to some taxonomic changes. For example, European Sphaerocarpos texanus plants differ from American S. texanus, both in their DNA sequences and in their spore characters, and so they are likely to be considered a separate species. Furthermore, European Sphaerocarpos michelii material includes three different forms based on spore characters; these are also confirmed by molecular research, and may be recognised at or below the rank of species.



Cargill, D.C. & J. Milne. 2013. A new terrestrial genus and species within the aquatic liverwort family Riellaceae (Sphaerocarpales) from Australia. Polish Botanical Journal 58(1): 71-80.

Douin R. 1907. Les Sphaerocarpus français. Revue Bryologique 34(6): 105-112.

Frye T.C. & L. Clark. 1937. Hepaticae of North America. University of Washington Publications in Biology 6: 105-113.

Haynes C.C. 1910. Sphaerocarpos hians sp. nov., with a revision of the genus and illustrations of the species. Bulletin of the Torrey Botanical Club 37(5): 215-230.

Haynes C.C. & M.A. Howe. 1923. Sphaerocarpales. North American Flora 14: 1-8.

Howe  M.A. 1899. The hepaticae and anthocerotes of California. Memoirs of the Torrey Botanical Club 7: 64-70.

Müller K. 1954. Die Lebermoose Europas. In: Rabenhorst’s Kryptogamenflora von Deutschland, Österreich und der Schweiz. 3. Auflage. Volume VI. Part 1. Leipzig, Akademische Verlagsgesellschaft Geest & Portig K.-G., Johnson Reprint Corporation (1971), New York, London.

Proskauer J. 1955. The Sphaerocarpales of South Africa. The Journal of South African Botany 21: 63-75.

Reimers H. 1936. Revision des europäischen Sphaerocarpus-Materials im Berliner Herbar. Hedwigia 76: 153-164.

Schill D.B., L. Miserere & D.G.Long. 2009. Typification of Sphaerocarpos michelii Bellardi, S. terrestris Sm. and Targionia sphaerocarpos Dicks. (Marchantiophyta, Sphaerocarpaceae). Taxon 58(2): 638-640.

Schuster R.M. 1992. Sphaerocarpales. In: The hepaticae and anthocerotae of North America V. Field Museum of Natural History, Chicago: 799-827.

Timme S.L. 2003. Sphaerocarpaceae. In: Bryophyte Flora of North America, Provisional Publication.


Sep 082016

One of the main problems with sampling largely from herbarium specimens, rather than from material that has been specifically collected for DNA work (rapidly dried in silica gel then maintained at low humidity), is that the quality of the DNA is unpredictable and usually rather poor. Therefore, despite starting out with 169 accessions and about 20 species of Marchantia, the actual successes, where we were able to get good quality DNA sequence data, were substantially lower. What we currently have is a slightly unbalanced data matrix, with 82 Marchantia accessions for rbcL, and 78 Marchantia accessions for psbA-trnH.

Reboulia hemisphaerica thallus, photographed by David Long (Long 34254)

Reboulia hemisphaerica thallus, photographed by David Long (Long 34254)

We also sequenced both rbcL and psbA-trnH from material of two accessions that we thought were Marchantia but where the sequences turned out to be Reboulia (from Texas) and Wiesnerella (from Bhutan). A quick check of the herbarium voucher specimens for both of these showed that they represented mixed collections of more than one complex thalloid species, for which the “wrong” plant parts had ended up in our silica dried tissue collection. Taking fortune from misfortune, both Reboulia and Wiesnerella form quite adequate outgroups for the phylogeny!

Wiesnerella denuda, photographed by David Long

Wiesnerella denuda thallus, photographed by David Long (Long 36267)

Out of the 20 species that we HAD hoped to sample, we ended up with only 12 named Marchantia species for rbcL (Marchantia polymorpha, M. paleacea, M. linearis, M. papillata, M. inflexa, M. emarginata, M. pinna, M. chenopoda, M. debilis, M. hartlessiana, M. quadrata and M. romanica), and 15 for psbA-trnH (Marchantia polymorpha, M. paleacea, M. linearis, M. papillata, M. inflexa, M. emarginata, M. pinna, M. chenopoda, M. debilis, M. globosa, M. pappeana, M. hartlessiana, M. subintegra, M. quadrata and M. romanica); we also had three Marchantia polymorpha subspecies (polymorpha, ruderalis and montivagans) and two Marchantia paleacea subspecies (paleacea and diptera).

That’s a little disappointing, representing, as it does, fewer than half of the 38 currently recognised species in the genus. However, we did also sequence a number of Marchantia accessions that had not been determined to species, and although many of them were good DNA matches to species that we had sampled, several are clearly different to everything else that we have included: one distinct lineage in Yunnan, China, another that occurs in Yunnan and Nepal, and a third in Indonesia and Malaysia. That’s balanced again by taxa that may not have been identified correctly; the psbA-trnH sequences from African material of M. debilis, M. globosa, M. pappeana and M. polymorpha, for example, are identical.

Intriguingly, in the “Preissia” clade, as well as M. romanica, there appear to be two lineages of Marchantia quadrata, one consisting of accessions from Denmark, Sweden and Sichuan, China, and the other with accessions from Svalbard, Norway and Utah, USA. These may tie in with subspecies quadrata (for the first lineage) and subspecies hyperborea (for the material from Svalbard and Utah), but the degree of genetic divergence is far higher than that found between many of the recognised species in Marchantia. It is a bit disconcerting, however, to notice that we have managed to overlook any Marchantia quadrata material from Scotland in our sampling!

The next step in the project, before it’s time to reveal any of the phylogenetic trees I’ve alluded to, is a phase of reciprocal illumination where we reconcile morphological information from the herbarium specimens with the information derived from the molecular sequence data. In other words, it’s time to double check our plant identifications, a part of the project that’s now in the capable hands of Dr David Long; the pile of Marchantia specimens is already on his desk!



Relevant posts

A rapid phylogeny of Marchantia, from the RBGE collections. I. Sampling

A rapid phylogeny of Marchantia, from the RBGE collections. II. Illuminating our sampling