An automated and containerized de novo assembly-based phylogenomics pipeline aimed to recover accurate and reproducible phylogenies from any combination of short reads and assemblies with particular emphasis on non-model taxa.
Orthogarden is a nextflow pipeline designed to leverage any combination of short reads and assemblies to generate a robust and accurate ML phylogeny with minimal user input. It attempts to accomplish this by first trimming reads, filtering reads for non-target contamination, de novo assembling reads, annotating assemblies, extracting orthologs from assemblies, and using harvested orthologs to create a phylogeny. A Nextflow-based architecture allows Orthogarden to run seamlessly from initiation to completion with little required knowledge of command line beyond installing dependencies and editing a config file to user standards. Extracting orthologs directly from de novo assemblies for direct comparison between taxa sets Orthogarden apart from other phylogenomics pipelines as it does not require a pre-selected suite of reference orthologs to function. Orthogarden is highly scalable and is demonstrated to generate accurate phylogenies from large and small datasets of varying sample quality.
Overview of pipeline:
nextflow (22.10.4+)
apptainer (1.1.8+)
git clone https://github.com/jacksonhturner/orthogarden.git
For more installation help, please see the wiki.
OrthoGarden requires a csv metadata file with the following headers id, r1, r2, ref, and augustus for each input sample. The id and augustus fields must be filled out for every sample and either both r1/r2 fields must be present OR the ref field for each sample.
Note
A helper script create_metadata.py has bin added to the bin directory and can assist in automating creation of metadata files for large directories of data.
Example metadata.csv:
id,r1,r2,ref,augustus
A_aegypti,,,/path/to/A_aegypti.fasta,aedes
A_albimanus,,,/path/to/A_albimanus.fasta,aedes
C_quinquefasciatus,C_quinquefasciatus_R1.fastq,C_quinquefasciatus_R2.fastq,,aedes
D_melanogaster,D_melanogaster_R1.fastq,D_melanogaster_R2.fastq,,fly
Note
The above example includes two samples using pre-assembled genomes (A_aegypti and A_albimanus) and two samples using paired-end reads (C_quinquefasciatus and D_melanogaster). Notice the Augustus references are allowed to vary.
To run the pipeline on a local linux server:
nextflow run /path/to/orthogarden/main.nf \
--input metadata.csv \
--threshold_val 0.9 \
--publish_dir results \
-profile local,two \
-resume
Note
This is a simplified usage script, for full details on all OrthoGarden parameters see the wiki/parameters.
Once you have apptainer and nextflow installed, to make sure the pipeline is configured correctly on your machine, run the following test from within the orthogarden directory:
mkdir -p ~/orthogarden_test
nextflow run main.nf \
--input tests/anopheles_pseudoref/pseudo_refs/metadata_test.csv \
--threshold_val 0.9 \
--publish_dir ~/orthogarden_test \
-profile local,two \
-resume
For more details on running the pipeline, installing prerequisites, or running on a slurm-based HPC, see the wiki.
The publish_dir contains all of the intermediate and final files produced by OrthoGarden runs. The work directory contains intermediate files (see note below). Files of particular interest are noted in the example publish results below.
Sample results directory:
.
├── publish
| ├── align_nt
| ├── augustus
| ├── design
| ├── iqtree
| | ├── run_iqtree
| | ├── run_iqtree.best_model.nex
| | ├── run_iqtree.best_scheme
| | ├── run_iqtree.best_scheme.nex
| | ├── run_iqtree.bionj
| | ├── run_iqtree.ckp.gz
| | ├── run_iqtree.contree
| | ├── run_iqtree.iqtree
| | ├── run_iqtree.log
| | ├── run_iqtree.mldist
| | ├── run_iqtree.model.gz
| | ├── run_iqtree.splits.nex
| | ├── run_iqtree.treefile 🌱
| | └── run_iqtree.ufboot
| ├── mafft
| ├── mstatx
| ├── mstatx_scores
| ├── orthofinder
| ├── orthofinder_finder
| ├── remove_thirds
| ├── summary
| ├── summary_table
| | ├── summary_table_with_genes.tsv 🌱
| | └── summary_table_with_taxon.tsv 🌱
| └── trimal
└── work
🌱 - Final treefile and relevant summary files.
flowchart TB
subgraph " "
subgraph params
v9["r1_adapter"]
v32["buffer_n"]
v28["limit_ogs"]
v25["ulimit"]
v42["retain_third_pos"]
v5["skip_qc"]
v27["threshold_val"]
v11["minimum_length"]
v0["input"]
v38["masking_threshold"]
v17["kraken_db"]
v8["skip_trim"]
v10["r2_adapter"]
end
v2([PARSE_METADATA])
v6([FASTQC_RAW])
v7([MULTIQC_RAW])
v12([CUTADAPT_ADAPTERS])
v14([FASTQC_TRIM])
v15([MULTIQC_TRIM])
v18([KRAKEN2])
v21([MEGAHIT])
v22([AUGUSTUS_FASTA])
v23([AUGUSTUS_READS])
v24([AUGUSTUS_PROT])
v26([ORTHOFINDER])
v29([ORTHOFINDER_FINDER])
v30([FIX_FRAMES])
v31([SUMMARY_TABLE])
v33([MAFFT])
v37([ALIGN_NT])
v39([TRIMAL])
v40([MSTATX])
v41([MSTATX_SCORES])
v43([REMOVE_THIRDS])
v44([IQTREE])
v45([IQTREE_WITH_THIRDS])
v0 --> v2
v2 --> v6
v6 --> v7
v2 --> v12
v9 --> v12
v10 --> v12
v11 --> v12
v12 --> v14
v14 --> v15
v17 --> v18
v2 --> v18
v2 --> v21
v2 --> v22
v21 --> v23
v22 --> v24
v23 --> v24
v24 --> v26
v25 --> v26
v24 --> v29
v26 --> v29
v27 --> v29
v28 --> v29
v29 --> v30
v29 --> v31
v32 --> v33
v29 --> v33
v32 --> v37
v33 --> v37
v30 --> v37
v32 --> v39
v37 --> v39
v38 --> v39
v32 --> v40
v39 --> v40
v40 --> v41
v32 --> v43
v39 --> v43
v43 --> v44
v39 --> v45
end
Note
If you are unfamiliar with Nextflow, the work directory consists of hexidecimal naming structure of directories with short two character names (e.g., "6f") containing one or more nested subdirectories with longer names (e.g., "19eeb79a9315d91d177d6fe985dc8f") that hold intermediate files, links, and Nextflow commands and logs. While this convention can be hard to understand, it is recommended to keep these files untouched until you are happy with your analysis, as they are used for Nextflow's resume functionality.