Juravel, K., Porras, L., Höhna, S., Pisani, D., & Wörheide, G. (2023). Exploring genome gene content and morphological analysis to test recalcitrant nodes in the animal phylogeny. PloS One, 18(3), e0282444.:
Exploring genome gene content and morphological analysis to test recalcitrant nodes in the animal phylogeny
Juravel, Ksenia 1; Porras, Luis 1; Höhna, Sebastian 1,2; Pisani, Davide 3; Wörheide, Gert 1,2,4, §
1 Department of Earth and Environmental Sciences, Paleontology & Geobiology, Ludwig-Maximilians-Universität München, Richard-Wagner-Str. 10, 80333 München, Germany
2 GeoBio-Center, Ludwig-Maximilians-Universität München, Richard-Wagner-Str. 10, 80333 München, Germany
3 School of Biological Sciences and School of Earth Sciences, University of Bristol, UK.
4 SNSB-Bayerische Staatssammlung für Paläontologie und Geologie, Richard-Wagner-Str. 10, 80333 München, Germany
§ corresponding author email: [email protected]
An accurate phylogeny of animals is needed to clarify their evolution, ecology, and impact on shaping the biosphere. Although datasets of several hundred thousand amino acids are nowadays routinely used to test phylogenetic hypotheses, key deep nodes in the metazoan tree remain unresolved: the root of animals, the root of Bilateria, and the monophyly of Deuterostomia. Instead of using the standard approach of amino acid datasets, we performed analyses of newly assembled genome gene content and morphological datasets to investigate these recalcitrant nodes in the phylogeny of animals. We explored extensively the choices for assembling the genome gene content dataset and model choices of morphological analyses. Our results are robust to these choices and provide additional insights into the early evolution of animals, they are consistent with sponges as the sister group of all the other animals, the worm-like bilaterian lineage Xenacoelomorpha as the sister group of the other Bilateria, and tentatively support monophyletic Deuterostomia.
This repository includes all the codes used to analyze the data for the various phylogenies.
All morphology related files can be found in the folder Morphology.
The results obtained in the different steps can be found in Morphology/Morphology_files.zip.
All genome gene content related files needed to reproduce the results can be found in the following folders:
Species_Files: The 47 proteomes needed as base data for all subsequent steps of dataset construction.
Code: all code/scripts needed to carry out the analyses, from dataset construction to phylogenetic analyses.
data_matrices_gene_content all data matrices analysed.
374 directories of genome gene content data that contain all the results from the different intermediate steps of dataset construction are too large to be provided here (uncompressed ~3 TB, compressed 645 GB) but can be provided upon request.
All the steps are listed and described below, for an overview see Figure 1.
Figure 1: Concise graphical illustration of the methodology and workflow used for the creation of the different datasets analysed (See the complete steps of gene content dataset creation for more detailed information.
- Extract proteins from genome based predictions for species of interest. 47 species were used here.
- Check for similarity all vs. all.
- Cluster the proteins into groups (homologous and orthologous).
- Extract ortho- and homogroups.
- Convert into matrices of absence and presence.
- Generate phylogeny.
- Statistical investigation.
Make sure to install the following tools:
OrthoFinder, MCL, DIAMOND, revBayes and homomcl.
(In folder Code)
OrthoFinder_for_MCL_after_DIAMOND.sh - Used to re-run OrthoFinder with different I- and E-values.
Nexus2Fasta.pl - Before pruning convert nexus format matrix back to fasta format data type.
Nexus_Pruning.sh - Used to prune out the species from the matrix.
does_specie_have_gene.py - Used to convert MCL output into fasta format data.
stat_from_bpcomp_tracecomp.sh - Used to check convergence for the 4 chains for all possible combinations to find the best parameters.
short_names_convert.sh - Used to translate from 4 letters short names of species to the formal names.
gene_content_original.Rev - code for revBayes adopted from willpett.
Create a folder with the complete set of proteomes for the species of interest. Moreover, follow the following steps:
Orthogroups
Run OrthoFinder with DIAMOND using different E-values.
Alternatively, run OrthoFinder once and just re-run DIAMOND on previously generated results for different E-values.
Use the pre-calculated results for MCL step using different I values.
To automate this step use the code similarly to the one in (adoption to your computer and settings of the system required) code for multiple OrthoFinder.
After generating the orthogroups file (Orthogroups.tsv, from OrthoFinder output) format it to fit the revBayes formatting in NEXUS using:
- tsv --> mcl output format
- mcl --> matrix of absence and presence
- matrix --> fasta format
- fasta --> NEXUS
(Or any other way to transfer the obtained results from tsv format to NEXUS).
-
Ensure to extract positions shared by less than two species (singletons extract).
-
Use the final output as input to the revBayse analyses.
Optional - 7. Extract convergence statistics for the phylogenies.
Optional - 8. Statistical investigation from MCMC samples, using the codes (Morphology, Genome gene content), plot with R script.
Optionally: To convert short format names of Species to full scientific names use Names_convert.sh.
Homogroups
For the homologs prediction after DIAMOND analysis step (BLAST all vs. all),extract length of the proteins for each species and run homomcl to create abc format file.
If you wish to have the final matrix use MCL.
If you want to analyze the different behaviors of the clusters use clmprotocols.
Repeat steps 1-7 from Orthogroups prediction with the output from MCL.
Outgroup sampling
In this part, the study was performed using the default parameters of all tools.
Repeat Homogroups and orthogroups prediction as described above but reduce the outgroups species, use
i) The complete taxon sampling (Opisthokonta);
ii) Ichthyosporea + Choanoflagellates + Metazoa (= Holozoa; dataset prefix Hol ), and
iii) Choanoflagellates + Metazoa (= Choanozoa; dataset prefix Cho)
Repeat the steps of Orthogroups and Homogroups with the new taxa combination proteomes.
Long-Branch Attraction (LBA)
We tested the effect of reducing specific taxa in two possible methods: first analyzing the data after reducing the taxa from start to end and second by reducing the data from an already pre-analyzed matrix. The last can significantly reduce the complexity of the analyses.
This part of the study was performed using the default parameters of all tools.
Repeat Homogroups and orthogroups prediction as described above but without the long-branched species
Caenorhabditis elegans, Pristionchus pacificus, and Schistosoma mansoni.
Also, combine this taxon sampling with Outgroup sampling section taxa sampling to create the ingroup test datasets - see Additional information.
Repeat the steps of Orthogroups and Homogroups with the new taxa combination proteomes.
Pruned
For datasets creation using the pruning method, use the initial 47 species dataset NEXUS and the script Nexus_Pruning.sh.
See file: "Genome gene content datasets protocol" .
To recreate the initial species dataset download all the files in this repository folder Species_Files ending with *.fasta.xz.
All NEXUS files are in Morphology_NEXUS.zip for the Morphology data and Run_2.zip.xz for the genome gene content second iteration (and in Run_1.zip.xz for the first iteration).
(In folder Tables)
Table S1: Information about all the data of species used in this research.
Table S2: The summary of all datasets settings and results for run 1.
Table S3: The summary of all datasets settings and results for run 2, a summary of all the details and the most probable tree for each of the 190 datasets tested, count of the support for each of the unique topologies observed by the individual posterior trees.
Section 3.1 - Row 4 to Row 73 Correspond to datasets in Supplementary Table 4. Section 3.2 - Row 74 to Row 193 Correspond to data for different E- and I- values tested (datasets for TPCT in Supplementary Figure 5) Section 3.3 - Row 194 to Row 405, Column C and Column AU Correspond to all statistical calculations in Supplementary Figure 4.
Table S4: Naming convention description for the long branch attraction tests for ingroups and outgroup-reduced datasets for long branches.
Table S5: Statistical hypothesis testing calculation results.
-
Select morphological datasets based on number of characters and taxon coverage of the 47 species of the genome gene content set.
-
Combine the character lists of the selected sets. Remove redundant and irrelevant characters.
-
Code two master matrices. One using Non-additive coding and another one using Reductive coding.
-
Produce another eight downsampled matrices for taxa exclusion experiments.
-
Analyze the ten matrices on Mr.Bayes v3.2 using the Mkv model (setting included in the matrices).
-
Statistical investigation (Statistical hypothesis testing calculation).
Homogroups - A set including homologous proteins that are predicted to be inherited from a common ancestor, all the proteins or parts, can include partial genes, orthologs, xenologs and paralogs. Contain any subset of the species, but no single species homogroups (proteins need to be shared by at least two species).
Orthogroups - A set of orthologous proteins that are predicted to be inherited from a common ancestor and separated by a speciation event, can also include in-paralogs and partial genes. Contain any subset of the species, but no single species homogroups (proteins need to be shared by at least two species).
Nchar - The number of positions in the final alignment (a matrix of alignment from 0 [absence] and 1 [presence]). Each character (column, if a matrix is species vs. chars) represents an orthogroup or homogroup (as defined previously) compared along with all the species.
Total Posterior Consensus Tree (TPCTree) - a majority rule consensus tree of all posterior trees from all converged chains of different Bayesian analyses. Two different TPCTrees were estimated: 1) based on an equal number of taxa independent from the methodology of parameters (TPCTree-all) and 2) based on homogroups or orthogroups methodology (TPCTree-partial).
Granulation - defined by the inflation parameter (I) in the MCL algorithm. It affects the cluster size, i.e., it defines the number of the predicted clusters for homogroups and orthogroups. The size of this parameter creates a scale, where small I values indicate fine-grained clustering, and large values a very coarse grained clustering. Increasing the I value leads to further splitting of the largest clusters, therefore smaller clusters.
Singleton - gene family which is coded as present in only a single species.
Naming convention in this work - The initial dataset was Opisthokonta (Opi) and contained data from genomes of 47 species. It was further divided into two subsets: a dataset with only Acoelomorpha (Aco, 44 species) and a dataset with Xenoturbella bocki alone (Xen, 41 species). Also, two additional datasets were created for the main Opi dataset and subsets (Aco and Xen), in which Fungi were excluded as outgroups Holozoa (Hol); and Choanozoa (Cho) where only species of the Choanoflagellates were included as outgroups. These are indicated with the three letters abbreviation of the outgroup sampling prefix and the suffix “dis” to distinguish from the next described dataset. Further, the subsets also excluded certain long-branched taxa in the ingroup and identified by the two letters suffix “ne” in the name of the dataset. The methodology type used for taxon reduction is indicated in the naming convention by “Ab'', for ab initio and “P” for pruning (see Supp. Table 5). For more details see Supp. Figure in Additional information and Genome gene content datasets protocol.
- for technical questions about genome gene content analyses or the statistical hypothesis testing: Ksenia Juravel ([email protected])
- for technical questions about morphology analyses: Luis Porras ([email protected])
- for all other questions the corresponding author: Gert Wörheide ([email protected])