skim2mito

skim2mito is a snakemake pipeline for the batch assembly, annotation, and phylogenetic analysis of mitochondrial genomes from low coverage genome skims. The pipeline was designed to work with sequence data from museum collections. However, it should also work with genome skims from recently collected samples.

Contents

• Setup
• Example data
• Input
• Output
• Filtering contaminants
• Assembly and annotation only
• Running your own data
• Getting help
• Citations

Setup

The pipeline is written in Snakemake and uses conda to install the necessary tools.

It is strongly recommended to install conda using Mambaforge. See details here https://snakemake.readthedocs.io/en/stable/getting_started/installation.html

Once conda is installed, you can pull the github repo and set up the base conda environment.

# get github repo
git clone https://github.com/o-william-white/skim2mito

# change dir
cd skim2mito

# setup conda env
conda env create -n skim2mito_env -f workflow/envs/conda_env.yaml
conda config --set channel_priority flexible

If you need to install the conda environment to a specific location, use the following example, where the prefix argument can be updated to include a specific path:

conda env create -n skim2mito_env --prefix /your_path/skim2mito_env -f workflow/envs/conda_env.yaml

Example data

Before you run your own data, it is recommended to run the example datasets provided. This will confirm there are no user-specific issues with the setup and it also installs all the dependencies. The example data includes simulated mitochondrial data from 25 different butterfly species.

To run the example data, use the code below. Note that you need to change the user email to your own address. The email is required by the Bio Entrez package to fetch reference sequences. The first time you run the pipeline, it will take some time to install each of the conda environments, so it is a good time to take a tea break :).

conda activate skim2mito_env

snakemake --cores 4 --use-conda --config user_email=user@example_email.com

Once this has finished, you can generate a snakemake report using the following command. As above, you need to change the user email to your own address.

snakemake --cores 4 --use-conda --config user_email=user@example_email.com --report skim2mito_report.html

Input

Snakemake requires a config.yaml and samples.csv to define input parameters and sequence data for each sample.

For the example data provided, the config file is located here config/config.yaml and it looks like this:

# path to sample sheet csv with columns for ID,forward,reverse,taxid,seed,gene
samples: config/samples.csv

# user email
user_email: user@example_email.com

# getorganelle reference (go_fetch, custom)
go_reference: go_fetch

# forward adapter
forward_adapter: AGATCGGAAGAGCACACGTCTGAACTCCAGTCA

# reverse adapter
reverse_adapter: AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT

# fastp deduplication (True/False)
fastp_dedup: True

# mitos refseq database (refseq39, refseq63f, refseq63m, refseq63o, refseq89f, refseq89m, refseq89o)
mitos_refseq: refseq39

# mito code (2 = Vertebrate, 4 = Mold, 5 = Invertebrate, 9 = Echinoderm, 13 = Ascidian, 14 = Alternative flatworm)
mitos_code: 5

# alignment trimming method to use (gblocks or clipkit)
alignment_trim: gblocks

# alignment missing data threshold for alignment (0.0 - 1.0)
missing_threshold: 0.5

# name of outgroup sample (optional)
# use "NA" if there is no obvious outgroup
# if more than one outgroup use a comma separated list i.e. "sampleA,sampleB"
outgroup: Eurema_blanda

# plot dimensions (cm)
plot_height: 20
plot_width: 20

The example samples.csv file is located here config/samples.csv and it looks like this (note that the seed and gene columns are only required if the custom getorganelle database option is specified in the config file):

Output

All output files are saved to the results direcotry. Below is a table summarising all of the output files generated by the pipeline.

Filtering contaminants

If you are working with museum collections, it is possible that you may assemble and annotate sequences from contaminant/non-target species. Contaminant sequences can be identified based on the blast search output or unusual placement in the phylogenetic trees (see blobtools and plot_tree outputs).

A supplementary python script format_alignments.py is provided to remove putative contaminants from alignments, and format the alignments for downstream phylogenetic analysis.

For example, let's say we wanted to remove all sequences from the sample "Kallima_paralekta" and atp6 gene sequences, you could run the script as shown below. The script works by identifying and removing sequences that have names with Kallima_paralekta or atp6 in the sequence names. The filtered alignments are written to a new output directory filter_alignments_output.

python workflow/scripts/format_alignments.py  \
   --input results/mafft_filtered/ \
   --cont Kallima_paralekta atp6 \
   --output filter_alignments_output

Note that the output fasta files have been reformatted so each alignment file is named after the gene and each sequence is named after the sample. This is useful if you would like to run our related pipeline gene2phylo for further phylogenetic analyses.

Assembly and annotation only

If you are only interested in the assembly of mitochondrial sequences and annotation of genes without the phylogenetic analysis, you can stop the pipeline from running the gene alignment and phylogenetic analyses using the --omit-from parameter.

snakemake --cores 4 --use-conda --config user_email=user@example_email.com --omit-from mafft

Running your own data

The first thing you need to do is generate your own config.yaml and samples.csv files, using the files provided as a template.

GetOrganelle requires reference data in the format of seed and gene reference fasta files. By default the pipeline uses a basic python script called go_fetch.py https://github.com/o-william-white/go_fetch to download and format reference data formatted for GetOrganelle.

go_fetch.py works by searching NCBI based on the NCBI taxonomy specified by the taxid column in the samples.csv file. Note that the seed and gene columns in the samples.csv file are only required if you want to provide your own custom GetOrganelle seed and gene reference databases.

You can use the default reference data for GetOrganelle, but I would recommend using custom reference databases where possible. See here for details of how to set up your own databases https://github.com/Kinggerm/GetOrganelle/wiki/FAQ#how-to-assemble-a-target-organelle-genome-using-my-own-reference

Getting help

If you have any questions, please do get in touch in the issues or by email o.william.white@gmail.com

Citations

If you use the pipeline, please cite our bioarxiv preprint: https://doi.org/10.1101/2023.08.11.552985

Since the pipeline is a wrapper for several other bioinformatic tools we also ask that you cite the tools used by the pipeline:

• Fastqc https://github.com/s-andrews/FastQC
• Fastp https://doi.org/10.1093/bioinformatics/bty560
• GetOrganelle https://doi.org/10.1186/s13059-020-02154-5
• Blastn https://doi.org/10.1186/1471-2105-10-421
• Minimap2 https://doi.org/10.1093/bioinformatics/bty191
• Blobtools https://doi.org/10.12688/f1000research.12232.1
• Seqkit https://doi.org/10.1371/journal.pone.0163962
• MITOS2 https://doi.org/10.1016/j.ympev.2012.08.023
• Gblocks (default) https://doi.org/10.1093/oxfordjournals.molbev.a026334
• Clipkit (optional) https://doi.org/10.1371/journal.pbio.3001007
• Mafft https://doi.org/10.1093/molbev/mst010
• Iqtree https://doi.org/10.1093/molbev/msu300
• ete3 https://doi.org/10.1093/molbev/msw046
• ggtree https://doi.org/10.1111/2041-210X.12628
• Multiqc https://doi.org/10.1093/bioinformatics/btw354