nsgtree

v0.5 March 3 2024

Why use NSGTree (over SGTree)

New Simple Genome Tree (NSGTree) is a computational pipeline for fast and easy construction of phylogenetic trees from a set of user provided genomes and a set of phylogenetic markers. NSGTree builds a species tree based on a concatenated alignment of all proteins that were identified with the provided markers HMMs. NSGTree also builds single protein trees for every marker protein that was found. It is a highly simplified version of SGTree. In contrast to SGTree, NSGTree does currently not support marker selection. Initial benchmarking showed that NSGTree is faster than SGTree (~20-50% depending on number of query faa and HMMs)

How to run it

Workflow

• Snakemake in a conda environment
• Input is a dir with faa files and a single file that contains a set of HMMs to identify makers

snakemake -j 24 --use-conda --config qfaadir="example" models="resources/models/rnapol.hmm" --configfile user_config.yml

or more general

snakemake -j <number of processes> --use-conda --config qfaadir="<dir with query faa>" models="<single file that contains all marker HMMs" --configfile <settings specified in user_config.yml>

• Alternatively, inputs can be a dir with query faa files and a dir with reference faa files, output dir will be created in the dir with query faa files

snakemake -j 24 --use-conda --config rfaadir="example_r"  qfaadir="example_q" models="resources/models/rnapol.hmm" --configfile user_config.yml

• Setting for alignment, trimming and tree building can be changed in user_config.yml
• Several sets of marker HMMs are provided in subdir resources/models/
• Results can be found in a subdir in <query faa dir>/nsgt_<analysis name>"

Apptainer / Shifter

• Apptainer: save the following into a run.sh script, specify path to dirs and config file as command line arguments, potentially edit further to specify different set of HMMs, adjust -j flag depending on available resources

LOCAL_FOLDER_WITH_QFAA_FILES=$1
LOCAL_FOLDER_WITH_RFAA_FILES=$2
USER_CONFIG=$3
apptainer run \
	docker://docker.io/fschulzjgi/nsgtree:0.4.3 \
	snakemake -j 16 \
	--snakefile "/nsgtree/workflow/Snakefile" \
	--use-conda \
	--config \
	qfaadir="$LOCAL_FOLDER_WITH_QFAA_FILES" \
	rfaadir="$LOCAL_FOLDER_WITH_RFAA_FILES" \
	models="/nsgtree/resources/models/rnapol.hmm" \
	--conda-prefix "/nsgtree/.snakemake/conda" \
	--configfile $USER_CONFIG

• Shifter on NERSC Perlmutter

shifterimg pull fschulzjgi/nsgtree:0.4.3

• Load the working directory that contains files with models, querydir and config file with shifter

shifter \
  --volume=$(pwd):/nsgtree/example \
  --image=fschulzjgi/nsgtree:0.4.3 \
  bash -c \
  "snakemake --snakefile /nsgtree/workflow/Snakefile \
  -j 24 \
  --use-conda \
  --config \
  qfaadir="/nsgtree/example/test" \
  models="/nsgtree/example/rnapol.hmm" \
  --conda-prefix /nsgtree/.snakemake/conda \
  --configfile /nsgtree/example/user_config.yml"

• Specify ref dir and query dir and config file with shifter, using the UNI56 models provided by nsgtree

qfaa=$1
rfaa=$2
configf=$3
shifter \
  --volume=$(pwd):/nsgtree/example \
  --image=fschulzjgi/nsgtree:0.4.3 \
  bash -c \
  "snakemake --snakefile /nsgtree/workflow/Snakefile \
  -j 24 \
  --use-conda \
  --config \
  qfaadir="/nsgtree/example/$qfaa" \
  rfaadir="/nsgtree/example/$rfaa" \
  models="/nsgtree/resources/models/UNI56.hmm" \
  --conda-prefix /nsgtree/.snakemake/conda \
  --configfile /nsgtree/example/$configf"

• For docker paths to modeldir, querydir, configfile can be loaded separately with the -v flag

docker pull fschulzjgi/nsgtree:0.4.3
docker run -t -i -v $(pwd):/nsgtree/modelsdir -v $(pwd)/test:/nsgtree/querydir --user $(id -u):$(id -g) fschulzjgi/nsgtree:0.4.3 snakemake --use-conda -j 16 --config qfaadir="querydir" models="/nsgtree/modelsdir/rnapol.hmm"

Acknowledgements

NSGTree was developed by the New Lineages of Life Group at the DOE Joint Genome Institute supported by the Office of Science of the U.S. Department of Energy under contract no. DE-AC02-05CH11231.