Animal phylogeny The animal kingdom comprises 34 large groups, or phyla - species within a phylum share a basic body plan, or set of characters that allows them to be grouped Relationships between phyla are difficult to determine, and remain controversial after 150 years of work - lack of homologous characters between phyla - simultaneous radiation of most groups during Cambrian hard to tell which groups are sister taxa, or who came first

Who is related to whom? MolluscAnnelidArthropod share obvious segmentation common ancestor, segmented?

Who is related to whom? MolluscAnnelidArthropod share trochophore larval stage common ancestor, with trochophore?

Molecular Phylogeny re-wrote the book Molluscs Annelids Arthropods Nematodes ancestor with trochophore larva ancestor with a cuticle Lophotrochozoa Ecdysozoa …Protostomes were divided into 2 major sub-groups Based on analysis of DNA, gene order…

Animal phylogeny: re-written by genetics Initial molecular phylogenetic studies compared genes that were highly conserved across phyla Compare a gene that is so important, it does not change much even over long periods of time - that way, creatures as different as a person and a sponge still have sequences that are similar enough to compare Small subunit ribosomal RNA gene (18S rRNA) is strongly conserved by selection; permits comparisons across very distantly related organisms

Long-Branch Attraction Problem Some taxa have fast-evolving DNA Often drop out at base of tree, clustered with: (a) primitive animals (b) other fast-evolvers, whom they may not be related to Early molecular studies found that nematodes dropped out near the cnidarians, suggesting they were basal bilaterians - this supported morphological analysis that said nematodes lacked a true coelom, so were basal to (more primitive than) other Bilaterians

Long-Branch Attraction Problem Some taxa have fast-evolving DNA Often drop out at base of tree, clustered with: (a) primitive animals (b) other fast-evolvers, whom they may not be related to This is an artifact (a false result) of how computer programs analyze DNA sequences, called long-branch attraction - sequences that are fast-evolving give very long branches on trees, which tend to “attract” other long branches sequences that are very different (fast mutating) get lumped together with other fast-evolving sequences

Long-Branch Attraction Problem Turned out that most nematodes have very fast-mutating DNA, so give long branches and tend to fall out near the base of any tree A slow-evolving nematode DNA sequence grouped with the arthropods, forming Ecdysozoa, the molting animals

not correcting for long-branch attraction correcting for long- branch attraction Tellford et al Curr. Biol. weird parasites often have long branches because they’ve gone thru lots of adaptive evolution

Tellford et al Curr. Biol. 1. Ctenophores move “in”, past sponges towards cnidarians - makes sense : they have tissues, gut; are cnidarian-like correcting for long- branch attraction

Tellford et al Curr. Biol. correcting for long- branch attraction 2. Acoel flatworms (never had a coelom) drop to base of Bilateria (outside Nephrozoa), away from more complex, long branch taxa - makes sense : no complex organs, gut, or coelom

Next-generation sequencing Recent animal phylogenies have relied on recent advances in sequencing technology and computational analysis 1) reverse-transcribe all mRNA in a tissue sample into DNA - this gives you a cDNA library of all the protein-coding genes that were being expressed in that tissue, which is called the transcriptome, or an EST library 2) use 454 pyrosequencing (“next-generation sequencing”) to get partial sequences for thousands of genes - you get many short sequences (~300 bp) which are hopefully overlapping, allowing computer to align them and infer the whole gene sequence from the pieces million nucleotides can be sequenced in 10 hr

Phylogenomics top represents the complete gene sequence, inferred from the aligned, overlapping pieces from individual runs each sequence is one piece of this particular gene, assembled by the computer based on the parts where they overlap (i.e., where the sequences are identical)

Phylogenomics Recent animal phylogenies have relied on recent advances in sequencing technology and computational analysis 3) computer algorithms are then used to piece together each gene sequence 4) the corresponding amino acid sequences are then used to compute the phylogenetic tree

Phylogenomics Problems with next-generation sequencing approaches to phylogenomics: - any given gene sequence may be incomplete - some genes may not be expressed in a given tissue, so they will not get sequenced - cost limits how many runs you can do, hence how many sequences you can obtain for a given species  lots of missing data in the final dataset: the sequence of any particular gene may be available for only some of the species you are trying to put into a phylogeny

Phylogenomics Problems with next-generation sequencing approaches to phylogenomics: - contamination from symbionts or food - determining whether copies of a gene from different taxa are orthologs, meaning copies of the same gene and not a related but different gene - some genes exits as families of similar genes, related by decent from one ancestral gene

Phylogenomics made-up example: say animals have Actin 1, Actin 2, and Actin 3 genes, descended from one ancestral Actin gene ancestral Actin gene actin actin actin gene duplication: 3 versions of actin were present in ancestral mollusc, with similar but different amino acids (present in 1 st bilaterian) GAFLSM.. GAFGSW.. TALLMM.. GAFLSM.. ancestral amino acids red letters = amino acids that changed by mutation since the 3 versions appeared in the ancestral mollusc

Phylogenomics made-up example: say there’s Actin 1, Actin 2, and Actin 3 genes in animals, descended from an ancestral Actin gene chiton squid clam snail GAFLSM.. MAFGSW.. TALLMM.. AAFPMM.. GWFGSP.. KRLLMY.. AAFPMM.. GAFLSP.. KRLLMQ.. 3 “Actin” genes in ancestral mollusc actin actin actin GAFLSM.. GAFGSW.. TALLMM.. diverge over time in each lineage related lineages should have more similar amino acids for each gene ortholog (copy of same original version) clam snail squid chiton

made-up example: say there’s Actin 1, Actin 2, and Actin 3 genes in animals, descended from an ancestral Actin gene chiton squid clam snail GAFLSM.. MAFGSW.. TALLMM.. AAFPMM.. GWFGSP.. KRLLMY.. AAFPMM.. GAFLSP.. KRLLMQ.. 3 “Actin” genes in ancestral mollusc actin actin actin GAFLSM.. GAFGSW.. TALLMM.. diverge over time in each lineage orthologs clam snail squid chiton paralogs same gene present in different species different versions of the same ancestral gene (in this case, of the original actin gene)

Phylogenomics Determining whether copies of a gene are orthologs - what you DON’T want: to align the copies of Actin 1 from squid, clam and chiton with Actin 2 from chiton !! GAFLSM.. MAFGSW.. TALLMM.. AAFPMM.. GWFGSP.. KRLLMY.. AAFPMM.. GAFLSP.. KRLLMQ.. actin actin actin this is a paralog of actin 1 gene: a divergent copy that exists alongside actin 1 in the genome chiton squid clam snail

Phylogenomics GAFLSM.. AAFPMM.. GAFLSP.. “actin 1” (or so you think..) really actin 2! chiton squid clam snail chiton squid snail clam at these 3 positions, “snail” now looks more related to chiton + squid because it has the same amino acids as they do! false information wrong phylogeny Determining whether copies of a gene are orthologs - what you DON’T want: to align the copies of Actin 1 from squid, clam and chiton with Actin 2 from chiton !!

Hejnol et al st large-scale phylogenomic study used 1,487 genes > 270,000 amino acid positions were analyzed complete data set could not even be fully analyzed due to requirement for huge amount of computer time - reduced 844-gene dataset used to test stability of proposed relationships

1st large-scale phylogenomic study used 1,487 genes - each gene only had to be present in 18 of the 94 included species (lots of missing data) - phoronid: only 2 genes used in analysis!

The rules of what makes a grouping “significant” (meaning we can really trust it is correct) state that you need bootstrap support of over 70% These are the numbers in black to the left of a node = 100% Of the major clades that group many phyla together (shaded boxes, or names on the left of the tree), in which can we be really confident? *

Based on the rules of what makes a grouping count as “significant” (meaning we can really trust it), only the following clades were secure: - Bilateria - Protostomia - Deuterostomia None of the other named clades of phyla I have been using were “significantly” supported

This is our current ‘consensus’ hypothesis of animal relationships polytomy: branching pattern remains unresolved; many possible sets of relationships equally likely, given available data still unresolved who is most closely related to whom for most of the phyla in Lophotrochozoa! “it’s a tie” = still controversial polytomy Tellford et al Curr. Biol.

Does it surprise you that after 100 years of effort, in the age of whole-genome sequencing, we still do not really know what phyla are sister to what other phyla?? What could be done to try to better resolve the relationships among phyla? What can you think of that would help to improve this tree either in terms of data or methods?