Digging Deeper: Population Dynamics Together With Individual-Based Fossil Phylogenies

In i of my last posts, I argued for looking out for ancestor-descendant relationships when putting upwards an evolutionary hypothesis based on fossil data. In this post, I volition explore a fleck the background of individual-based phylogenies together with why nosotros should proceed inward heed population processes, when analysing such information sets.

One way to avoid having direct ancestors inward our dataset, together with other issues of palaeophylogenetic studies such every bit problematic coding of characters, is to non utilisation composite or higher taxa. Instead of (fossil) species together with genera, i scores each private or local population every bit an operational taxonomic unit of measurement (OTU). In my fourth dimension every bit professional person scientist together with blogger, I dealt amongst ii of such information sets: our Osmundaceae together with Osmundales matrices (Bomfleur, Grimm & McLoughlin 2015, 2017; run across also this post) together with the information matrix of Tschopp, Matteus & Benson (2015) on "long-necks" (see this post), the dinosaurs familiar to us at to the lowest degree since Littlefoot's adventures. Influenza A virus subtype H5N1 third, plesiosaurian "sea monsters" (Fischer et al. 2018), came to my attending via Twitter, which belike volition live theme of roughly other ship service on the Genealogical World of Phylogenetic Networks.

A graph (neighbour-net) depicting morphological variety inward the rhizomes of Rex ferns together with relatives (order Osmundales). We (phylognetic classification) highlighted past times colour.

 

For taxonomic purposes – the Definition of biologically meaningful units comprising similar together with closely related individuals such every bit fossil genera – such elaborate matrices are a huge bound forwards of our agreement of the past. Palaeontological nomenclature is oftentimes overly specific, together with I spare you lot (and me) a give-and-take of species concepts inward palaeontology. There's picayune dot to theoretically hash out something fifty-fifty neontology has no universal answer to, but a working practical i (Mallet 1995; for those that are into it, read Mallet 2010). To me, it's simple, organisms that cast a coherent grouping of (putative or likely) relatives inward infinite together with time, should receive got a collective name. If they are indistinguishable, when in that location is evidence for largely overlapping features (as e.g. inward the instance of the ancestors of the western Eurasian Common Beech, Fagus sylvatica s.l., Denk 1999), i should live careful assigning unlike species epithets.

Organisms that are component subdivision of the same evolutionary lineage, should receive got a collective name, too.

In palaeontology, nosotros tend to give (or are forced to past times our peers during review) rattling similar things unlike names when nosotros discovery them inward unlike places or stratigraphic units (time zones). Facing ii principal problems:

1. Diagnoses that are non mutually exclusive. Influenza A virus subtype H5N1 mutual phenomenon for plants, e.g. inward the above-mentioned Rex ferns, aggravated past times the fact that works life fossils are unremarkably found every bit disarticulated, dispersed organs. 
2. Diagnoses so specific that whatever other fossil must direct stand upwards for a unlike species (not rare inward instance of animals, particularly vertebrates). Especially when they are good preserved together with present a lot of traits, every bit e.g. a fully preserved skeleton or completely preserved works life organ.

Palaeontological nomenclature (not always, but often). Fossils of the same (ancient, instantly extinct) species, the Blue Cow (see below) present unlike sets of preserved characters, hence, are named differently, i fifty-fifty every bit unlike genus. They may live fused only when all iii fossil species would live found at the same fossil site together with inward the same strata (and fifty-fifty then, non without resistance).

(Mis)Guided past times theoretically fixed but impractical species concepts, palaeontologists receive got used missing information to fence for unlike names. In extreme cases, this has Pb authors to assign a blossom to i genus, the seeds inward it (i.e. in-situ) to another, together with the pollen found in-situ to a third, ending upwards describing several novel species based on a unmarried fossil. Why? Because the same seeds together with pollen are found isolated (dispersed), together with nosotros can't live for certain that they come upwards from the same species (or genus) that produced the blossom i was lucky to find. This avoids beingness incorrect (in a way), because you lot don't receive got to brand a call. But also renders palaeo-taxonomy a completely useless endeavour. If nosotros desire to delineate a lineage through infinite together with time, nosotros remove names that dot us to the right fossils. If nosotros give a thing a name, that call should receive got a purpose (and meaning).

For phylogenetic inference these matrices impose nevertheless roughly novel problems. They inevitably receive got a higher per centum of missing information when nosotros seek to actually accept inward whatever is possible. For our (character-limited, nosotros could score simply 45) Osmundales rhizomes, a 5th of the matrix cells are blank (i.e. enquiry marks). Influenza A virus subtype H5N1 swiss cheese. In the instance of the 10-times larger (character-wise) matrix of Tschopp et al. two-thirds are blank, a cheesy aerosol, together with a fleck to a greater extent than than one-half inward the 5-times larger (than ours) matrix of Fischer et al. Influenza A virus subtype H5N1 special missing information bias subcategory is lack of overlap inward grapheme suites. Imagine, you lot only discovery fossil Influenza A virus subtype H5N1 together with C of the Blue Cow, the Not-Blue Cow together with the Blue Cow-oid, but non fossil B, the Blu-ish Cow that tin connect the both. There would live a rattling high gamble to interpret the shared bits of bluish-ness of the Cow-oid every bit convergences, together with house it far away of the (Blue) Cow clade, its actual species.

Even nether improve circumstances, missing information tin touching on phylogenetic inference. Best-case scenario is that it reduces the determination capacity of the optimisation algorithm. Worst-case scenario that it leads to inferring clades that hit non reverberate a monophyletic group (sensu Hennig).

Let's assume nosotros receive got all iii Blue Cow fossils together with a well-preserved fossil of Red Cows, the sis species of the Blue Cow. For our Blu Cow problem, the next would hand off during tree inference (very probable when using parsimony or distance-based inferences, together with peradventure when using a probabilistic framework, maximum likelihood or Bayesian inference).

Topological risks when using individual-based trees. An utterly incorrect tree would live reconstructed because of missing information together with pathetic information overlap. The iii representatives of Blue Cows are artificially separated, together with the sis species, the Red Cow, misplaced.

Branch support, unremarkably non rattling high to start with, may decrease further; topological ambiguity tin increase. This is a practical problem, which nosotros tin solve past times reducing the matrix to the best-preserved OTUs to infer a backbone tree or network, together with and so embed the poorer-preserved OTUs into the phylogeny past times associating them amongst the found groups or clades past times additional agency (such every bit sheer similarity, every bit e.g. done past times Tschopp et al. together with us).

The alter of branch back upwards when using information subsets (less taxa, less characters) is also a skillful indicator for missing information bias. Which brings us to under-utilised bootstrapping together with jack-knifing resampling procedures. For the cows, the bootstrap consensus network (see Schliep et al. 2017, opened upwards access, for an introduction) could inform us that the colour-less moo-cow is either a Red or Blue Cow and, hence, should live excluded from a tree inference, together with that Parabos is non a skillful sis genus, but a unopen relative, potential sister, or fifty-fifty co-specific amongst the Blue Cow.

Another number is that the to a greater extent than honest (unfiltered) together with comprehensive a matrix is, the less treelike volition live its signal (see Osmundales graph above, which essentially lacks tree-like portions). Most morphological traits receive got been evolved to a greater extent than than once, independently or inward parallel together with at unlike fourth dimension together with roughly tin live pretty variable already at the population, intraspecific or interspecific flat — past times using individuals, nosotros encompass both (see our Osmundales dataset together with papers). For lineages amongst living representatives (as inward the instance of the Osmundales) this tin live quantified or at to the lowest degree qualified, a molecular phylogeny tin live used to position tree-compatible together with tree-incompatible morphological traits. For extinct lineages, nosotros are essentially crossing marginal sea H2O ice during polar nighttime hoping the candle inward our manus holds all the way. Unless nosotros receive got e.g. a marine invertebrate amongst nearly continuous fossil records providing thousands of individuals together with reflecting transition inward fourth dimension together with infinite at fine scale.

Using the Osmundales approach on Tschopp et al.'s long-necks matrix. Colouration of dots betoken whether the taxon would acquit good or non inward tree-inference. Taxa amongst picayune missing information together with under-average iDV (individual Delta Value; run across Göker & Grimm 2008, opened upwards access, for an explanation) are favourable (green dots), those amongst pathetic coverage together with over-average iDV volition deed every bit "rogue" OTUs, i.e. trim back branch back upwards together with increment topological ambiguity.

Individuals are snapshots

It is crucial to realise this: An private (fossil or living) is non only the production of the evolutionary history of the grouping but also of ± reticulate population processes inside a grouping of closest-relatives (populations, species, species aggregates, genera). Reticulation – interbreeding, introgression, hybridisation – is much to a greater extent than mutual than to a greater extent than oftentimes than non assumed, also inward animals (an example fresh from the press).

But it goes deeper than this. Just give the next a thought. We utilisation an inference method that assumes that a biological (composite) unit, a lineage or species, ever splits into ii (lineages or species) that together with so evolve independently from each other for a primary biological unit, an individual, that – inward the instance of sexual reproduction – combines ii lineages, namely that of the manlike individual nurture together with that of the mother.

And fatherly together with motherly inherited traits (genetic or morphological) are gratis to mix inward subsequent populations, together with this mixing decides what the species – the amount of interacting individuals together with populations – looked similar at a for certain fourth dimension inward a particular place. And speciation is slow (see this bright newspaper past times back 150 Ma or more, depending on the definition), nosotros receive got something known every bit "hybrid vigor": the hybrid (often allopolyploids) volition live something intermediate or novel (phenotypically) together with rattling productive.

Also for (still) living animals, genetics furnish plenty of evidence of reticulation, today together with inward the past. Just accept the bear information nosotros re-used for our Intertwining trees together with networks paper (Schliep et al. 2017).

A consensus network showing all alternatives for inter-species relationships inward bears. The network includes iii molecular trees based either on nuclear autosomal introns (ncAI, inherited from both parents), the consummate Y-chromosome (YCh, inherited from the fathers) or a consummate mitochondrial factor laid (inherited from the mothers). The initiatory off modern bears living almost v Ma agone (see Heath, Huelsenbeck & Stadler 2014, fig. 10, providing a novel molecular dating method: the "fossilized birth-death" dating), were non also concerned almost species boundaries. Spoiler: the well-resolved sis human relationship of Brown together with Polar Bear is partly due to their (obviously) frequent crossing until today. Genetic-cladistically, the Polar Bear is a Brown Bear.

Influenza A virus subtype H5N1 few to a greater extent than things to proceed inward heed inward this context: Professional serpent breeders (an instance what you lot tin buy) brand a fortune past times breeding obscure hybrids betwixt serpent lineages that must receive got been isolated 100 or to a greater extent than i thou 1000 years: today they alive on unlike continents separated past times wide, former oceans. Lizards are notoriously ignorant of species barriers Mallet. And nosotros shoot Grizzly (a Brown Bear variant/ subspecies) – Ice Bear hybrids to proceed the species apart together with ensure the survival of the Ice Bear (while melting the poles). Natural hybrids of ii highly specialised bear species every pocket-size fry tin distinguish, together with whatever vertebrate zoologist simply past times component subdivision of the skeleton. The unnatural inter-continental hybrid snakes are unremarkably infertile, but a pup of an occassional dear thing betwixt a Grizzly together with an Ice Bear is not. The "Grolar Bear" is also featured inward Memorandum], who was a peachy fan of Republic of Finland together with (from the rattling beginning) neutral evolution (his miss translated Motoo Kimura's book into German). Like belike most (or all) population geneticists, he did non concord so much regard for Hennigian phylogenetics or Farrisian cladistics, which he considered unnatural frameworks. Already retired, when I started studying genetics (because it was much less attempt for me than other things), he gave a population genetic course of written report inward the semester breaks, unremarkably attended past times roughly few Finnish (summer) invitee students, together with i year, ii Finns were joined past times me (I did most of my Diplom-biology courses inward the break, because during the semester I was busy amongst my actual studies: geology-palaeontology; dorsum then, inward the usual relaxed way amongst getting on every bit much excursions every bit possible).

Prof. Sperlich's experimental set-up instruction you lot the basics of neutral development together with population genetics/phenetics.

If you lot repeat the experiment oftentimes enough, you lot volition eventually receive got ruby-red moo-cow populations, together with so a ruby-red moo-cow species. As you lot tin imagine, the smaller the starting population amongst the mutation (and a novel phenotype), the to a greater extent than probable it volition live that the ruby-red cows accept over. [This was simply the start, later the 2-week course, you lot knew everything (for roughly while) i needs to know almost the principles of population genetics together with phenetics, mostly those of fruit flies, because of their extremely fast generation turnover.]

Naturally this takes time, together with if a species is large together with its populations already a fleck isolated, the novel trait may ready only inward roughly component subdivision of the species. The evolutionary history (and genetics) of the beech provides such examples. When isolation becomes permanent, nosotros may receive got i sibling species amongst no change, together with roughly other carrying the novel trait. There may live to a greater extent than traits scattered inward the ancestral species, which are together with so passed to roughly but non all descendants. But amongst fossil histories, fourth dimension is what nosotros receive got plenty of. And Mother Nature is permanently throwing coins.

In the (very) rare case, a mutation is beneficial, choice steps inward together with massively accelerates the process. So every bit long every bit the population amongst the private having the beneficial mutation is non isolated, the novel beneficial (= to a greater extent than offspring/dominant inward heterozygotes) phenotypic (or genotypic) characteristic volition spread through the entire species: Influenza A virus subtype H5N1 novel species emerges from the hollow musical rhythm out of an old, similar a butterfly from a cocoon.

Let's add together this spice to the instance I used inward my earlier post. 

My theoretical instance uses ii basic observations:

1. Isolated lineages volition accumulate unique derived features (illustrated past times colouring gradients together with novel colours). Either because of cladogenesis (a lineage splits into two), e.g. the differentiation of the ancestral species into its geographic "races", forming 4 species: violet, still-black splitting inward lush together with pale green, grey, together with yellow) or neutral development (e.g. blueish into dark blue, xanthous into low-cal orange).
2. Already established lineages coming into contact may intrograde when unopen relatives (direct sisters, e.g. the violet together with purple) or hybridise when less close, but sharing a mutual ancestry (blue together with red, both surviving descendants of the violet together with imperial sisters, cast pink). 

Now nosotros blank out in i lawsuit to a greater extent than roughly of the populations ...

... together with reconstruct a phylogenetic tree including all individuals assuming morphology mirrors phylogeny, together with that nosotros receive got no missing information issues or confusing convergences – no homoplasy (which nosotros unremarkably receive got a lot when using morphological data).

Because of development – organisms differentiate and/or alter given fourth dimension inward various ways (dichotomously, linear or horizontal), together with at unlike stair – together with inevitable gaps inward fourth dimension together with space, our individual-level tree includes clades that oftentimes are right (green branches) or non alone incorrect (yellow branches) together with roughly that are (red branches).

With real-world information the gamble for inferring partly or alone incorrect clades increases.

• Long-branch attraction would ensure that X*1 would live placed every bit sis to whatever of the other most derived (asterisks = modern) individuals, because it is also distinct from its actual ancestor Z2. The As may live drawn closer to the root, because of their full general similarity to the non-derived Zs. As effect the B- together with C-comprising clades would travel sisters (I telephone outcry upwards this phenomenon "short-branch culling"). 
• The cross S1 may deed every bit rogue OTU, or clit its putative sis S2 to the blueish B-clade away from the T together with U clades. Influenza A virus subtype H5N1 tree needs to create upwards one's heed what has to a greater extent than weight: joining S1 amongst the Bs (then S2 volition move, too) or bring together S2 inward a STU-clade (then S1 volition stay). 
• One of the most frequent branching artefacts inward morphology-based trees (and occasionally, inward molecular trees) is that derived, long-branched taxa are placed every bit early on diverged sisters to their (much) less diverged relatives (sisters or ancestors). Likely inevitable for our instance inward instance of the B+P clade (monophyletic grades – paraphyletic clades paradox).

For comparison, the evolutionary tree depicting correctly all relationships betwixt the sampled individuals (and their populations/species).

Quite a unlike thing.

Why palaeontologists should proceed inward heed modern-day dynamics

There is currently no method to reconstruct such an evolutionary tree. Since nosotros are looking at needles separated past times many generations, nosotros may live tempted discarding population together with speciation processes shaping together with messing upwards morphologies (phenotypes). But populations are oftentimes non homogenous (morphologically together with genetically). And non all populations of a species are necessarily homogenised. They eventually be, but nosotros receive got no command almost the dot nosotros tapped into the puddle amongst our fossil individual. Was is during a speciation process, was it long later isolation together with trait sorting took place? The tinier together with detailed the puzzle pieces, nosotros utilisation inward our matrix, the to a greater extent than nosotros receive got to bargain amongst a mix of lineage-sorted (phylogenetically controlled) together with highly volatile, single-tree-incompatible inter-/intra-species phenotypic characters.

The to a greater extent than complex an organism gets, the to a greater extent than flickly it becomes inward selecting its breeding partners/ measures, together with the to a greater extent than dynamic (volatile) together with various its manifesting species. And so fifty-fifty geographically restricted species/populations tin receive got quite a complex population structure, every bit latterly shown past times plant; worst-case: a wind-pollinated such every bit the beech. The papers past times Mallet instructed me to halt pondering almost species concepts. Still, the currently living beech species are slow to circumscribe (morphologically together with genetically) but are patently non the production of a strictly dichotomous development — a trivial serial of cladogeneses every bit reflected inward the commonly seen phylogenetic trees.

When you lot await at a palaeo-phylogenetic papers (individual or composite taxon based), the authors unremarkably piece of job nether the next assumptions:

1. Species are stable, fixed biological primary units that don't alter over time.
2. Species tin live characterised past times morphology.
3. Speciation is a strictly dichotomous process (a serial of cladogenesis events); whenever in that location is a visible morphological change, nosotros receive got ii novel (sister) species, together with in that location has been a cladogenesis event.
4. Morphology mirrors phylogeny (often: the most-parsimonious solution for the morphological information reflects the truthful tree).
5. The improve papers realised: Genera are no skillful units; a phylogeny should live ideally based on individuals together with non on (potentially artificial) composite taxa — in that location are ii sides to this medallion (see above).

When you lot receive got dealt inward low-level evolution, today together with into the past times (see e.g. our studies on the beech trees), i.e. intra-specific together with inter-specific relationships, it is clear that:

1. Species are non stable, fixed biological units. Neither morphologically nor genetically.
2. Morphology tin out-run molecular systematics, when it comes to define species (coherent-as-possible, genetic together with morphology-wise, but performs worse when it comes to infer inter-species relationships (or higher up: phylogenies). 
3. Reticulation is non rare. Species barriers are permeable together with volition live broken, to a greater extent than so inward plants but also inward animals. Influenza A virus subtype H5N1 dominion of pollex is, the to a greater extent than primitive the animal, the to a greater extent than permeable the species boundaries. The only enquiry is: volition this receive got an impact on what nosotros see? Also, speciation tin live a complex physical care for (see doodles higher upwards together with the example of the beech).
4. Morphological development is non parsimonious in general.
   
5. Genera (at to the lowest degree inward plants) are unremarkably skillful units for tracing a grouping inward the past. Because they tin live defined inward a much to a greater extent than stable way, so that they don't vanish inward a blink of fourth dimension (in the instance of beech: roughly sixty Ma, inward the instance of modern Osmundaceae: 200 Ma). 

So, what tin I do?

Given the fragmentation of the fossil record, nosotros receive got no (simple) way to determine how much of the (individual) morphological variety is the outcome of lineage-sorting, the phylogeny of the grouping nosotros written report — (true) tree-compatible signals; together with how much relates to stochastic, potentially reticulate low-level differentiation processes — tree-incompatible signals. And in that location is to a greater extent than enemy fire: Parallel or convergent development of the same traits inward related or unrelated lineages, epigenetic phenomena leading to phenotypes non fixed past times genotypes but reacting to the electrical flow environment. Using morphology alone, nosotros also receive got no objective way to define species or higher taxa (such every bit genera) to filter lineage-specific traits.

Hence, we remove to explore the signals inward our information amongst methods that allow visualising conflict (e.g. back upwards consensus networks, Schliep et al. 2017) or tin bargain amongst incompatible signals (e.g. the quick-and-easy to do "Neighbor-Net", Bryant & Moulton 2002, 2004). Not simply infer a tree. Most relationships seen inward those trees may non live (entirely) wrong, but all (unfiltered) morphological information sets present conflict. At best, the inferred tree(s) capture(s) aspects, but they don't capture evolutionary history at large. Less-so using individual-based matrices than the classic ones using composite taxa, because they (the individual-based onces) are data-wise much improve (usually less biased past times concepts). And closer to the complex reality.

Long-known, past times the way. When the (then, mostly a) theory of development was soundless young, researchers realised that signals from morphology are riddled past times internal conflict, which tin live visualised using networks but non trees.

A network based on St George Mivart's 1865 together with 1867 primate trees. Depicted inward D. Morrison's Is this the initiatory off network from conflicting datasets? ship service (GWoN, September 2012)

Links
See the link listing of my ship service on Where receive got all the ancestors gone? for a listing of links for farther reading.

In The challenging together with puzzling ordinary beech – a history, I state the story how nosotros reconstructed the evolutionary history of the beech. It's no miracle to pose upwards holistic hypothesis almost how a genus (or whatever other group) unfolds inward infinite together with time, simply a articulation effort.

Cited literature
Bomfleur B, Grimm GW, McLoughlin S. 2015. Osmunda pulchella [now: Osmundastrum pulchellum] sp. nov. from the Jurassic of Sweden—reconciling molecular together with fossil evidence inward the phylogeny of modern royal ferns (Osmundaceae). BMC Evolutionary Biology 15:126.
Bomfleur B, Grimm GW, McLoughlin S. 2017. The fossil Osmundales (Royal Ferns)—a phylogenetic network analysis, revised taxonomy, together with evolutionary classification of anatomically preserved trunks together with rhizomes. PeerJ 5:e3433.
Bryant D, Moulton V. 2002. NeighborNet: an agglomerative method for the construction of planar phylogenetic networks. In: Guigó R, together with Gusfield D, eds. Algorithms inward Bioinformatics, Second International Workshop, WABI. Rome, Italy: Springer Verlag, Berlin, Heidelberg, New York, p. 375-391.
Bryant D, Moulton V. 2004. Neighbor-Net: An agglomerative method for the construction of phylogenetic networks. Molecular Biology together with Evolution 21:255-265.
Denk T. 1999. The taxonomy of Fagus in western Eurasia together with the ancestors of Fagus sylvatica s.l. Acta Palaeobotanica Supplement 2:633-641.
Fischer V, Benson RBJ, Druckenmiller PS, Ketchum HF, Bardet N. 2018. The evolutionary history of polycotylid plesiosaurians. Royal Society Open Science 5:17217 [e-pub].
Göker M, Grimm GW. 2008. General functions to transform associate information to host data, together with their utilisation inward phylogenetic inference from sequences amongst intra-individual variability. BMC Evolutionary Biology 8:86.
Grimm GW, Denk T. 2010. The reticulate origin of modern airplane trees (Platanus, Platanaceae) - a nuclear mark puzzle. Taxon 59:134-147. 
Heath TA, Huelsenbeck JP, Stadler T. 2014. The fossilized birth–death physical care for for coherent calibration of divergence-time estimates. Proceedings of the National Academy of Sciences 111:E2957–E2966.
Mallet J. 1995. A species Definition for the Modern Synthesis. Trends inward Ecology together with Evolution 10:294-299.
Mallet J. 2001. The speciation revolution. Journal of Evolutionary Biology 14:887-888.
Mallet J. 2007. Hybrid speciation. Nature 446:279-283.
Mallet J. 2008. Hybridization, ecological races, together with the nature of species: empirical evidence for the ease of speciation. Philosophical Transactions of the Royal Society of London, Series B 363:2971–2986.
Mallet J. 2010. Why was Darwin’s persuasion of species rejected past times twentieth century biologists? Biology together with Philosophy 25:497-527.
Schliep K, Potts AJ, Morrison DA, Grimm GW. 2017. Intertwining phylogenetic trees together with networks. Methods inward Ecology together with Evolution DOI:10.1111/2041-210X.12760.
Tschopp E, Mateus O, Benson RBJ. 2015. Influenza A virus subtype H5N1 specimen-level phylogenetic analysis together with taxonomic revision of Diplodocidae (Dinosauria, Sauropoda). PeerJ 3:e857.