1. more info on prepTG

prepTG creates a database directory of genomes to search for homologous instances of reference/query gene-clusters in using fai.

The input is simply a directory of either FASTA or GenBank formatted files - with CDS features for the latter - representing bacterial genomes or metagenomes.

For eukaryotic genomes full GenBank format with CDS features are expected; however, FASTA formatted assemblies may instead be provided if a "reference proteome" is provided.

Check out example commands for prepTG on the 4. basic usage examples wiki page.

Gene-calling bacterial genomes using prodigal and pyrodigal

For bacterial genomes or bacterial metagenomes, users are able to use pyrodigal (default) or prodigal to perform de novo gene calling. More recently, we also have the availability of prodigal-gv as an option for gene-calling when phages are the gene clusters of interest.

Gene-mapping in eukaryotic genomes using miniprot

For eukaryotic genomes, users are able to map a high-quality gene-calling prediction for some reference genome to the remainder of the genomes. This approach is generally recommended only for single-species investigations and has only been tested with microbial eukaryotic genomes of a modest size (e.g. fungii, not gigantic genomes such as those of plants).

prepTG usage

Usage: prepTG [-h] [-i INPUT_DIR] [-g GTDB_TAXON] [-gr GTDB_RELEASE] [-d DOWNLOAD_PREMADE] -o OUTPUT_DIR [-r] [-ro] [-l LOCUS_TAG_LENGTH]
              [-gcm GENE_CALLING_METHOD] [-m] [-rp REFERENCE_PROTEOME] [-cst] [-ma] [-c THREADS] [-mm MAX_MEMORY] [-v]

        Program: prepTG
        Author: Rauf Salamzade
        Affiliation: Kalan Lab, UW Madison, Department of Medical Microbiology and Immunology

        Prepares a directory of target genomes for being searched for query gene clusters using fai.

        Premade databases of representative genomes are available for the following genera:

        Acinetobacter (n=1,643), Bacillales (n=3,150), Corynebacterium (n=726), Enterobacter (n=878),
        Enterococcus (n=937), Escherichia (n=2,436), Klebsiella (n=1,022), Listeria (n=353),
        Mycobacterium (n=744), Pseudomonas (n=2,666), Salmonella (n=308), Staphylococus (n=496),
        Streptomyces (n=1,555), Streptococcus (n=2,452), Cutibacterium (n=27), Neisseria (n=414),
        Lactobacillus (n=541), and Micromonospora (n=211).

        In addition, users can simply request all genomes belonging to a specific species/genus
        in GTDB R214 to be downloaded.

        ----------------------------------------------------------------------------------------
        > Example commands:

        1. Setup a prepTG database which includes some local genomes in FASTA format

            $ prepTG -i User_Genomes_Directory/

        2. Setup a prepTG database which includes some local genomes and all Cutibacterium granulosum
           genomes in GTDB R214:

           $ prepTG -i User_Genomes_Directory/ -g "Cutibacterium granulosum" -o prepTG_Database/

        3. Setup local prepTG database by downloading a premade one of representative
           Cutibacterium genomes:

           $ prepTG -d Cutibacterium -o prepTG_Database/

        ----------------------------------------------------------------------------------------
        > Considerations
        If FASTA format is provided, assumption is that genomes are prokaryotic and
        pyrodigal/prodigal will be used to perform gene-calling. Eukaryotic genomes can
        be provided as FASTA format but the --reference-proteome file should be used in
        such case to map proteins from a reference proteome (from the same species ideally)
        on to the target genomes. This will prevent detection of new genes in gene-clusters
        detected by fai but synchronize gene-calling and allow for better similarity
        assessment between orthologous genes.

        If you are interested in inferring horizontal gene transfer using salt downstream
        and are working with bacterial genomes - consider issuing the "--mge-annotation"
        flag to annotate phage, plasmid and IS element associated proteins.

        If GenBank files are provided, CDS features are expected and further each CDS
        feature must contain a "translation" qualifier which features the protein sequence
        and optionally a "locus_tag" qualifier. Options to consider if not every CDS has a
        "locus_tag" include --rename-locus-tags and --rename-problem-gbk-lts.

Options:
  -h, --help            show this help message and exit
  -i, --input-dir INPUT_DIR
                        Directory with target genomes (either featuring GenBanks or FASTAs).
  -g, --gtdb-taxon GTDB_TAXON
                        Name of a GTDB valid genus or species to incorporate genomes from.
                        Should be surrounded by
                        quotes (e.g. "Escherichia coli").
  -gr, --gtdb-release GTDB_RELEASE
                        GTDB release to use. [Current default is R220].
  -d, --download-premade DOWNLOAD_PREMADE
                        Download and setup pre-made databases of representative genomes
                        for specific taxon/genus. Provide name of the taxon,
                        e.g. "Escherichia"
  -o, --output-dir OUTPUT_DIR
                        Output directory, which can then be provided as input for the
                        "-tg" argument in fai.
  -r, --rename-locus-tags
                        Whether to rename locus tags if provided for CDS features in
                        GenBanks.
  -ro, --rename-problem-gbk-lts
                        Whether to rename locus tags of only problem GenBank
                        files which have CDS but no locus_tag qualifier. By default
                        such GenBank files are skipped.
  -l, --locus-tag-length LOCUS_TAG_LENGTH
                        Length of locus tags to set. Default is 3, allows for <~18k
                        genomes.
  -gcm, --gene-calling-method GENE_CALLING_METHOD
                        Method to use for gene calling. Options are: pyrodigal, prodigal,
                        or prodigal-gv. [Default is pyrodigal].
  -m, --meta-mode       Flag to use meta mode instead of single for pyrodigal/prodigal.
  -rp, --reference-proteome REFERENCE_PROTEOME
                        Provide path to a reference proteome to use for protein/
                        gene-calling in target genomes - which should be in FASTA
                        format.
  -cst, --create-species-tree
                        Use skani to infer a neighbor-joing based species
                        tree for the genomes.
  -ma, --mge-annotation
                        Perform MGE annotation of proteins - for
                        bacterial genomes only.
  -c, --threads THREADS
                        The number of threads to use.
                        [Default is 1].
  -mm, --max-memory MAX_MEMORY
                        Uses resource module to set soft memory limit. Provide
                        in Giga-bytes. Configured in the shell environment
                        [Default is None; experimental].
  -v, --version         Get version and exit.

Detailed Usage on Major Options:

Option (Short)	Option (Long)	Description
-i	--input-dir	Directory with target genomes (either featuring GenBanks or FASTAs). If GenBank files are provided - they must feature CDS features with gene predictions. They should preferably also include a "locus_tag" for each CDS feature, but this can be overcome using the --rename-locus-tags or --rename-problem-gbk-lts options.
-d	--download-premade	Download and setup pre-made databases of representative genomes for specific taxon/genus selected by skDER. For details on available databases, see this wiki page. Provide the name of the taxon, e.g. "Escherichia".
-g	--gtdb-taxon	Name of a GTDB valid genus or species to incorporate genomes from. Should be surrounded by quotes (e.g. "Escherichia coli").
-gr	--gtdb-release	GTDB release to use. [Current default is R220].
-o	--output-dir	Output directory, which can then be provided as input for the -tg argument in fai.
-l	--locus-tag-length	Length of locus tags to set. Default is 3, allows for <~18k genomes.
-r	--rename-locus-tags	Whether to rename locus tags for CDS features if GenBank files are provided as input.
-ro	--rename-problem-gbk-lts	Whether to rename locus tags of only problem GenBank files which have CDS but no locus_tag qualifier. By default such GenBank files are skipped.
-gcm	--gene-calling-method	Method to use for gene calling. Options are: pyrodigal, prodigal, or prodigal-gv. [Default is pyrodigal].
-m	--meta-mode	Flag to use metagenomic mode for pyrodigal/prodigal.
-rp	--reference-proteome	Provide path to a reference proteome to use for miniprot-based protein mapping in target genomes. The input proteome should be in FASTA format.
-cst	--create-species-tree	Use skani ANI predictions to infer a neighbor-joining based species tree for the genomes. Not recommended if working with diverse genome sets spanning multiple species.
-ma	--mge-annotation	Perform MGE annotation of proteins - for bacterial genomes only. This is only used for the salt program downstream, be careful when using for large datasets as it increases prepTG runtime significantly.