DNA Methylase Finder

DNA Methylase Finder is a tool to detect/predict DNA methylase genes of prokaryotes. More specifically, it will detect genes with DNA methylase DOMAINS by querying their amino acid sequences. BUT WHY?

☠️ There are many types of methylases, not just DNA methylases (i.e. RNA methylases)

👻 Common gene annotation tools like prokka, and even PGAP, usually label genes as "methylase", therefor the annotation specificity is ambiguous

👾 Many genes with DNA methylase domains have other domains (i.e. helicases or nucleases), and the annotation of the other domain is used as the gene label from prokka and PGAP.

💥 DNA methylation in prokaryotes is proving to be interesting, especially as new technologies enable surveying of 6mA, 5mC, and 4mC methylation marks/motifs. Understand the genes responsible for these marks is important..

This tool was able to correctly call all "gold standard" DNA methlyases in the REBASE database (proteins with M. prefix). De novo discovery of methylases was conducted in hundreds of bacterial genomes and metagenomic assemblies. I have not noticed any systematic false positive patterns, but false negative rate would be difficult to assess.

Input

Nucleotide contigs/genomes (.fna) or amino acids (.faa).

Output (important files)

All input types:

.DNA_methylases.combined.summary.tsv : summary table of all detected DNA methylases, their coordinates, subtype, and specificity. Open in text editor or Excel, etc.

.DNA_methylases.combined.faa : multi-fasta of amino acid sequences of predicted DNA methylase genes (including merged fragments)

Nucleotide inputs only:

neighborhood_annotations/*neighbors.gb : DNA methylase gene neighborhood map of DNA methylase and surrounding genes. Good for determining if each DNA methylase gene is part of a Restriction Modification System. Open in GenBank file/plasmid viewer, e.g. SnapGene Viewer or UGENE.

Schematic

Installation instructions

NOTE: I have only tested this on Linux and MacOS (See extra installation instruction for MacOS)

1. Make sure you have conda installed

conda -V

2. Clone this repo

git clone https://github.com/mtisza1/DNA_methylase_finder.git

3. Change to the DNA_methylase_finder directory:

cd DNA_methylase_finder

4. Create the conda environment called dna_methylase_finder:

conda env create --file dna_methylase_finder.yml

NOTE: if you can't use conda, you could probably install the dependencies manually with little issue (listed in .yml file). Python 3 required.

5. If you are using MacOS, you need a few additional packages to make it work:

conda activate dna_methylase_finder

conda install -c conda-forge -c bioconda sed grep findutils coreutils

6. Download and unpack the databases (3.5 Gb compressed, 11 Gb decompressed):

# you should still be in the DNA_methylase_finder directory
wget https://zenodo.org/record/6647341/files/DNA_methylase_finder_DBS_v1.0.tar.gz 
tar -xvf DNA_methylase_finder_DBS_v1.0.tar.gz
rm DNA_methylase_finder_DBS_v1.0.tar.gz

versions of packages/tools that this tool was tested on

bcbio-gff v0.6.9
bioawk v1.0
biopython v1.78
blast v2.12
hmmer v3.3.2
python v3.9.7
prodigal-2.6.3

Usage

1. Activate the conda environment:

conda activate dna_methylase_finder

2. Run the python wrapper (specify the path to DNA_methylase_finder when runnning). Probably best not to do runs from within the repo directory.

python DNA_methylase_finder/run_DNA_methylase_finder.py -it nucl -f MY_CONTIGS.fna -r MY_CONTIGS_DMF1 -t 16

Run test AA file (distributed with repo) from Bacteroides fragilis to check everything is working:

python DNA_methylase_finder/run_DNA_methylase_finder.py -it AA -f DNA_methylase_finder/test_seqs/mtases_GCA_016889925.1.faa -r TEST_AAs

This test should return hits for all 13 sequences.

Summary table example

gene name	sub-type	contig	start position	end position	orientation	motif guess	AAI to top hit	AF to top hit
SRS022713_1554_3_	Type_III	SRS022713_1554	3503	5047	-	not_found	not_found	not_found
SRS022713_1833_25_	Type_II	SRS022713_1833	27134	28228	-	RAATTY	99.73 AAI	99.73 AF
SRS022713_4879_20_	Type_I	SRS022713_4879	21121	22674	+	GAYNNNNNNNTAYG	88.91 AAI	99.23 AF
SRS022713_4903_2_	Type_II	SRS022713_4903	431	1150	+	not_found	not_found	not_found
SRS022713_511_9_	Type_II	SRS022713_511	15370	16641	+	not_found	not_found	not_found
SRS022713_5642_14_	Type_IIG	SRS022713_5642	16963	20844	-	not_found	not_found	not_found
SRS022713_5668_33_	Type_I	SRS022713_5668	37289	38794	+	not_found	not_found	not_found
SRS022713_5789_1_	Type_II	SRS022713_5789	4	2946	-	not_found	not_found	not_found
SRS022713_6589_13_	Type_II	SRS022713_6589	14021	14851	-	not_found	not_found	not_found
SRS022713_6589_18_	Type_II	SRS022713_6589	18841	19410	-	GATC	99.47 AAI	99.47 AF
SRS022713_7606_6_	Type_II	SRS022713_7606	8461	9021	+	GATC	91.94 AAI	99.47 AF
SRS022713_7854_31_	Type_II	SRS022713_7854	26974	27720	-	GATC	82.66 AAI	99.6 AF
SRS022713_7854_34_	Type_II	SRS022713_7854	29433	30503	-	not_found	not_found	not_found
SRS022713_8242_47_	Type_I	SRS022713_8242	42714	44282	+	not_found	not_found	not_found
SRS022713_870_47_	Type_II	SRS022713_870	56482	57627	+	not_found	not_found	not_found
SRS022713_130_5_@SRS022713_130_7_#merged	Type_II	SRS022713_130	56482	57627	+	not_found	not_found	not_found
SRS022713_7801_31_@SRS022713_7801_32_#merged	Type_I	SRS022713_7801	56482	57627	+	not_found	not_found	not_found
SRS022713_7936_6_@SRS022713_7936_7_#merged	Type_IIG	SRS022713_7936	56482	57627	+	not_found	not_found	not_found

Tips

• Setting --neighborhoods False will reduce the runtime by quite a bit but no neighborhood maps will be generated. It is the suggested setting for scanning large databases for DNA methylase genes.
• Increasing number of CPUs available with -t will make the tool run faster.
• See something weird? Open an Issue on this repo describing your problem in detail.
• A large majority of prkaryotic genes with DNA methylase domains methylate genomic DNA with a particular motif specificity. Some of these genes with not be "active" at all times. A small number of these genes are not involved in motif-specific methylation (they could be involved in DNA repair).

Help Menu

usage: run_DNA_methylase_finder.py [-h] -it INPUT_TYPE -f INPUT_FILE -r RUN_TITLE [--version] [-t CPU]
                                   [--meth_hmms METHYLASE_HMMS] [--cdd_plus_hmms CDD_PLUS_HMMS]
                                   [--legit_domains LEGIT_DOMAIN_LIST] [--motif_blastp MOTIF_ANNOTATE_BLASTP]
                                   [--subtype_hmms SUBTYPE_ANNOTATE_HMM] [--prod_args PROD_ARGS] [--pid PID] [--cov COV]
                                   [--s_subunit_hmms S_SUBUNIT_HMM] [--re_hmms RE_HMM] [--neighborhoods NEIGHBORHOODS]
                                   [--merge MERGE]

DNA Methylase Finder v1.0.1

options:
  -h, --help            show this help message and exit

 REQUIRED ARGUMENTS for DNA Methylase Finder, v1.0.1:
  -it INPUT_TYPE, --input_type INPUT_TYPE
                        OPTIONS: nucl, AA -- nucl PREFERRED! nucl is a nucleotide fasta file .fna extension. Each header
                        must be unique before the first space character. AA is an amino acid fasta file with fasta file
                        .fna extension. Each header must be unique before the first space character.
  -f INPUT_FILE, --input_file INPUT_FILE
                        nucl file with .fna extenstion, prodigal directory, or AA seq file with .faa extension.
  -r RUN_TITLE, --run_title RUN_TITLE
                        Name of this run. A directory of this name will be created. Must be unique from older runs or
                        older run will be renamed. Must consist of ONLY letters, numbers and underscores (_)

 OPTIONAL ARGUMENTS for DNA Methylase Finder, v1.0.1:
  --version             show program's version number and exit
  -t CPU, --cpu CPU     Default: 4 -- Number of CPUs available for run.
  --meth_hmms METHYLASE_HMMS
                        Default: standard database -- Hmmer-formatted file of HMMs of putative DNA methylases
  --cdd_plus_hmms CDD_PLUS_HMMS
                        Default: standard database -- Hmmer-formatted file of HMMs of all CDD + putative DNA methylases
  --legit_domains LEGIT_DOMAIN_LIST
                        Default: standard list -- text file (1 entry per line) with names of DNA methyalse Hmmer models
  --motif_blastp MOTIF_ANNOTATE_BLASTP
                        Default: standard database -- BLASTP-formatted file of all REBASE DNA methylase proteins with
                        motif tag
  --subtype_hmms SUBTYPE_ANNOTATE_HMM
                        Default: standard database -- Hmmer-formatted file of HMMs of Olveira subtype HMMs with subtype
                        tag@@
  --prod_args PROD_ARGS
                        Default: -c -p meta -- arguments for prodigal in quotation marks. Only relvant for --input_type
                        nucl (-it nucl). Make sure to keep settings to produce AA, nucleotide, and gtf files from prodigal
                        step. Do not use memory or CPU arguments.
  --pid PID             Default: 80 -- minimum threshold for AA Percent Identity of predicted methylase gene to REBASE
                        homolog to predict motif specificity
  --cov COV             Default: 80 -- minimum threshold for alignment coverage of predicted methylase gene to REBASE
                        homolog to predict motif specificity
  --s_subunit_hmms S_SUBUNIT_HMM
                        Default: standard database -- Hmmer-formatted file of HMMs specificity subunit proteins
  --re_hmms RE_HMM      Default: standard database -- Hmmer-formatted file of HMMs restriction enzyme (endonuclease)
                        proteins
  --neighborhoods NEIGHBORHOODS
                        Default: True -- Make DNA methylase gene neighborhood maps? True - OR - False
  --merge MERGE         Default: True -- Merge adjacent DNA methylases with the assumption that they are a broken ORF?
                        True - OR - False