TideCluster

TideCluster is a software tool designed to identify tandem repeats in genome assemblies by utilizing TideHunter to detect tandem repeats clustering these repeats based on similarity using mmseqs2 and NCBI BLAST. The software runs in four steps as
outlined below:

Tidehunter step: In this initial step, Tidehunter is utilized to identify tandem repeats. As TideHunter's performance diminishes with larger sequences, the input fasta file is divided into smaller overlapping segments, with each segment analyzed individually. Results from individual segments are parsed and merged into a single GFF3 file. Tandem repeats detected in this step are often fragmented into multiple
overlapping pieces.

Clustering step: Prior to clustering, all arrays that do not meet the minimum length requirement are removed from the analysis and saved in a separate GFF3 file. Arrays exceeding the minimum length requirement are clustered based on similarity. Clustering occurs in two stages. First, mmseqs2 is employed in the initial round of clustering. The second round involves an all-to-all comparison using NCBI-BLAST, followed by graph-based clustering. The GFF3 file from the Tidehunter step is updated to include cluster assignment information. Simple sequence repeats are excluded from the clustering step and are analyzed separately.

Annotation step: Consensus sequences from TideHunter for each cluster are examined by RepeatMasker against a library of tandem repeats. The resulting annotation for each tandem repeat is used to update the information in the GFF3 file. This step is optional.

TAREAN step: In this final step, the Tandem Repeat Analyzer (TAREAN) estimates consensus sequences using a k-mer-based approach on the original sequences from the
reference. Consensus sequences of simple sequence repeats are evaluated separately,
as TAREAN performs poorly on tandem repeats with short monomers. The results of the
analysis are saved in an HTML summary.

TRC Consensus Clustering and Superfamily Analysis Consensus sequences from TAREAN are clustered to identify similary or nearly identical TRCs tha were not clustered together initially in clustering step. This clustering groups the TRCs into higher-level clusters referred to as Superfamilies, based on sequence similarity.

K-Mer Interval Tandem Repeat Estimation (KITE) The K-Mer Interval Tandem Repeat Estimation (KITE) method is applied to analyze individual Tandem Repeat Arrays (TRAs) within each TRC. In contrast to TAREAN, which provides monomer size estimates for the entire TRC, KITE also estimates monomer sizes for each array by evaluating the distances between all repetitive k-mers across the tandem repeat array, with each tandem array analyzed individually. This method facilitates the detection of higher-order repeats and captures the variability in monomer size across different tandem repeat arrays. The KITE analysis includes:

Monomer size estimate for the entire TRC : This estimate may differ from the TAREAN estimate due to the different methodologies employed.
Monomer size estimates for individual TRAs : These estimates are provided as:
- Primary estimate : The estimate with the highest score.
- Alternative estimates : Estimates with lower scores, these estimates help in identifying higher-order repeats.


TideCluster workflow scheme

Installation

TideCluster is available on Anaconda repository. To install TideCluster run we recommend to install it using Mamba an extremely fast replacement for the Conda package manager

In case you do not have Mamba installed, you can install it using conda to your base environment:

conda install -n base -c conda-forge mamba

Then install TideCluster using Mamba:

mamba create -n tidecluster -c conda-forge -c bioconda -c petrnovak tidecluster

If you are getting Illegal instruction error message from TideHunter, try to compile TideHunter from source and replace TideHunter conda binary with newly compiled binary. In activated TideCluster environment run:

wget https://github.com/yangao07/TideHunter/releases/download/v1.4.3/TideHunter-v1.4.3.tar.gz
tar -zxvf TideHunter-v1.4.3.tar.gz && cd TideHunter-v1.4.3
make
cp bin/TideHunter $CONDA_PREFIX/bin

Usage

usage: TideCluster.py [-h] [-v] {tidehunter,clustering,annotation,tarean,run_all} ...

Wrapper of TideHunter
    This script enable to run TideHunter on large fasta files in parallel. It splits
    fasta file into chunks and run TideHunter on each chunk. Identified tandem repeat 
    are then clustered, annotated and representative consensus sequences are extracted.
    
     
    

positional arguments:
  {tidehunter,clustering,annotation,tarean,run_all}
                        TideHunter wrapper
    tidehunter          Run wrapper of TideHunter
    clustering          Run clustering on TideHunter output
    annotation          Run annotation on output from clustering stepusing reference library of tandem repeats
    tarean              Run TAREAN on clusters to extract representative sequences
    run_all             Run all steps of TideCluster

options:
  -h, --help            show this help message and exit
  -v, --version         show program's version number and exit

    Example of usage:
    
    # first run tidehunter on fasta file to generate raw GFF3 output
    TideCluster.py tidehunter -c 10 -f test.fasta -pr prefix 
    
    # then run clustering on the output from previous step to cluster similar tandem repeats
    TideCluster.py clustering -c 10 -f test.fasta -pr prefix  -m 5000
    
    # then run annotation on the clustered output to annotate clusters with reference
    # library of tandem repeats in RepeatMasker format
    TideCluster.py annotation -c 10 -pr prefix -l library.fasta
    
    # then run TAREAN on the annotated output to extract representative consensus
    # and generate html report
    TideCluster.py  tarean -c 10 -f test.fasta -pr prefix
    
    Recommended parameters for TideHunter:
    short monomers: -T "-p 10 -P 39 -c 5 -e 0.25"
    long monomers: -T "-p 40 -P 3000 -c 5 -e 0.25" (default)
    
    For parallel processing include -c option before command name. 
    
    For more information about TideHunter parameters see TideHunter manual.
    
    Library of tandem repeats for annotation step are sequences in RepeatMasker format
    where header is in format:
    
    >id#clasification

TideHunter step

usage: TideCluster.py tidehunter [-h] -f FASTA -pr PREFIX [-T [TIDEHUNTER_ARGUMENTS]] [-c CPU]

options:
  -h, --help            show this help message and exit
  -f FASTA, --fasta FASTA
                        Path to reference sequence in fasta format
  -pr PREFIX, --prefix PREFIX
                        Base name for output files
  -T [TIDEHUNTER_ARGUMENTS], --tidehunter_arguments [TIDEHUNTER_ARGUMENTS]
                        additional arguments for TideHunter in quotes, default value: -p 40 -P 3000 -c 5 -e 0.25)
  -c CPU, --cpu CPU     Number of CPUs to use

Clustering step

usage: TideCluster.py clustering [-h] -f FASTA [-m MIN_LENGTH] -pr PREFIX [-g GFF] [-nd] [-c CPU]

options:
  -h, --help            show this help message and exit
  -f FASTA, --fasta FASTA
                        Reference fasta
  -m MIN_LENGTH, --min_length MIN_LENGTH
                        Minimum length of tandem repeat, default (5000)
  -pr PREFIX, --prefix PREFIX
                        Prefix is used as a base name for output files.If --gff is not provided, prefix will be also usedto identify GFF file from previous tidehunter step
  -g GFF, --gff GFF     GFF3 output file from tidehunter step. If not provided the file named 'prefix_tidehunter.gff3' will be used
  -nd, --no_dust        Do not use dust filter in blastn when clustering
  -c CPU, --cpu CPU     Number of CPUs to use

Annotation step

usage: TideCluster.py annotation [-h] -pr PREFIX [-g GFF] [-cd CONSENSUS_DIRECTORY] -l LIBRARY [-c CPU]

options:
  -h, --help            show this help message and exit
  -pr PREFIX, --prefix PREFIX
                        Prefix is used as a base name for output files.If --gff is not provided, prefix will be also usedto identify GFF3 file from previous clustering step
  -g GFF, --gff GFF     GFF3 output file from clustering step. If not provided the file named 'prefix_clustering.gff3' will be used
  -cd CONSENSUS_DIRECTORY, --consensus_directory CONSENSUS_DIRECTORY
                        Directory with consensus sequences which are to be annotated. If not provided the directory named 'prefix_consensus' will be used
  -l LIBRARY, --library LIBRARY
                        Path to library of tandem repeats
  -c CPU, --cpu CPU     Number of CPUs to use

LIBRARY is fasta file with tandem repeats reference library. Required format for sequence names in fasta is >seqid#class/subclass

TAREAN step

usage: TideCluster.py tarean [-h] [-g GFF] -f FASTA -pr PREFIX [-c CPU]

options:
  -h, --help            show this help message and exit
  -g GFF, --gff GFF     GFF3 output file from annotation or clustering stepIf not provided the file named 'prefix_annotation.gff3' will be used instead. If 'prefix_annotation.gff3' is not
                        found, 'prefix_clustering.gff3' will be used
  -f FASTA, --fasta FASTA
                        Reference fasta
  -pr PREFIX, --prefix PREFIX
                        Prefix is used as a base name for output files.If --gff is not provided, prefix will be also usedto identify GFF3 files from previous clustering/annotation step
  -c CPU, --cpu CPU     Number of CPUs to use
  -M MIN_TOTAL_LENGTH, --min_total_length MIN_TOTAL_LENGTH
                        Minimum combined length of tandem repeat arrays within a single cluster, required for inclusion in TAREAN analysis.Default (50000)

Run all steps

usage: TideCluster.py run_all [-h] -f FASTA -pr PREFIX [-l LIBRARY] [-m MIN_LENGTH] [-T [TIDEHUNTER_ARGUMENTS]] [-nd] [-c CPU]

options:
  -h, --help            show this help message and exit
  -f FASTA, --fasta FASTA
                        Reference fasta
  -pr PREFIX, --prefix PREFIX
                        Base name used for input and output files
  -l LIBRARY, --library LIBRARY
                        Path to library of tandem repeats
  -m MIN_LENGTH, --min_length MIN_LENGTH
                        Minimum length of tandem repeat (5000)
  -T [TIDEHUNTER_ARGUMENTS], --tidehunter_arguments [TIDEHUNTER_ARGUMENTS]
                        additional arguments for TideHunter in quotes, default value: -p 40 -P 3000 -c 5 -e 0.25)
  -nd, --no_dust        Do not use dust filter in blastn when clustering
  -c CPU, --cpu CPU     Number of CPUs to use
  -M MIN_TOTAL_LENGTH, --min_total_length MIN_TOTAL_LENGTH
                        Minimum combined length of tandem repeat arrays within a single cluster, required for inclusion in TAREAN analysis.Default (50000)

Example for full pipeline

TideCluster.py tidehunter -c 40 -pr cen6_sat -f CEN6_ver_220406.fasta 
TideCluster.py clustering -c 40 -pr cen6_sat -f CEN6_ver_220406.fasta
TideCluster.py annotation -c 40 -pr cen6_sat -l library.fasta
TideCluster.py tarean -c 40 -pr cen6_sat -f CEN6_ver_220406.fasta

# or simpli run all steps at once
TideCluster.py run_all -c 40 -pr cen6_sat -f CEN6_ver_220406.fasta -l library.fasta

Output

Tidehunter Step

prefix_tidehunter.gff3 - GFF3 file with tandem repeats detected by TideHunter.
prefix_chunks.bed - BED file showing how the reference sequence was split into chunks for parallel processing.

Clustering Step

prefix_clustering.gff3 - GFF3 file with tandem repeats identified by mmseqs2 and BLASTN. Tandem repeat regions in the GFF3 file are labeled by Tandem Repeat Cluster ID (TRC1, TRC2, etc.). Each TRC is described by the repeat_type attribute. repeat_type can be either TR (Tandem Repeat) or SSR (Simple Sequence Repeat).
prefix_clustering.gff3_1.gff3 - Intermediate file with tandem repeats clustered by mmseqs2.
prefix_consensus - Directory with consensus sequences for each cluster as identified by TideHunter. There is one FASTA file per cluster. Each FASTA file contains all consensus sequences identified by TideHunter for a given cluster.
prefix_consensus_1 - Intermediate directory with consensus sequences for each cluster as identified by mmseqs2.

Annotation Step

prefix_annotation.gff3 - GFF3 file with tandem repeats annotated by RepeatMasker. Annotations are shown as additional attributes in the GFF3 file.
prefix_annotation.tsv - Summarized annotation for each TRC cluster in a tab-delimited format.

TAREAN Step

prefix_index.html - Main HTML report, other reports are linked from this file.
prefix_tarean_report.html - HTML report with tandem repeat annotations.
prefix_tarean_report.tsv - File with tandem repeat annotations in a tab-delimited format.
prefix_tarean - Directory containing subdirectories with detailed TAREAN output for each TRC cluster.
prefix_consensus_dimer_library.fasta - FASTA file with consensus sequences for each TRC cluster. This sequences can be used as a library for similarity based annotation using RepeatMasker. This file is created only for TRC clusters that pass the minimum combined length threshold.

Updating gff3 file based on manual annotation

If you want to update GFF3 file with manual annotation, you can use tc_update_gff3.py script. This script will update "Name" attribute in GFF3 based on the conversion table. Conversion table is tab-delimited file with two columns. First column is original value of Name attribute and the second column is a new value.

usage: tc_update_gff3.py [-h] -g GFF3 -t TABLE -o OUTPUT [-a ATTRIBUTE_NAME]

Update gff3 attributes based on conversion table.

options:
  -h, --help            show this help message and exit
  -g GFF3, --gff3 GFF3  gff3 file
  -t TABLE, --table TABLE
                        Conversion table as tab-delimited file. First column if original attribute value, second column is new attribute value.
  -o OUTPUT, --output OUTPUT
                        output file gff3
  -a ATTRIBUTE_NAME, --attribute_name ATTRIBUTE_NAME
                        attribute name to update, default attribute is "Name"

Reannotation of tandem repeats using similarity based approach

TideCluster.py tarean step produces FASTA file with representative sequences for each TRC which can be used as library for similarity based annotation using RepeatMasker. This library file is available in prefix_consensus_dimer_library.fasta. Similarity based annotation can fill gaps in annotation of tandem repeats provided by TideCluster/TideHunter. To reannotate tandem repeats using RepeatMasker, you can use tc_reannotate.py script. This script will run RepeatMasker using TRC library and then filter RepeatMasker output to retain only high-quality TRC hits. Overlapping TRC hits are merged and regions shorter than twice the monomer length are excluded from the output.

Usage:

usage: tc_reannotate.py [-h] (-r REPEATMASKER_FILE | -s REF_SEQ) -f FASTA_FILE [-c CPU] -o OUTPUT [-d]

options:
  -h, --help            show this help message and exit
  -r REPEATMASKER_FILE, --repeatmasker_file REPEATMASKER_FILE
                        RepeatMasker output file
  -s REF_SEQ, --ref_seq REF_SEQ
                        FASTA file to annotated by TRC library
  -f FASTA_FILE, --fasta_file FASTA_FILE
                        Fasta file wiht TRC library used for RepeatMasker search
  -c CPU, --cpu CPU     Number of CPUs to use
  -o OUTPUT, --output OUTPUT
                        GFF3 output file
  -d, --debug           Keep temp files for debuging

Credits

TideCluster utilizes Tidehunter [https://github.com/Xinglab/TideHunter] for tandem repeat detection and TAREAN for reconstruction of consensus sequences of tandem repeats. If you use TideCluster please cite:

https://github.com/kavonrtep/TideCluster
TAREAN: a computational tool for identification and characterization of satellite DNA from unassembled short reads (https://doi.org/10.1093/nar/gkx257)
TideHunter: efficient and sensitive tandem repeat detection from noisy long-reads using seed-and-chain (https://doi.org/10.1093/bioinformatics/btz376)

Authors

Petr Novak, Jiri Macas, Laboratory of Molecular Cytogenetics, Biology Centre CAS, Czech Republic

License

GNU General Public License v3.0

Name		Name	Last commit message	Last commit date
Latest commit History 90 Commits
tarean		tarean
LICENSE		LICENSE
README.md		README.md
TideCluster.py		TideCluster.py
changelog.md		changelog.md
credits.html		credits.html
requirements.txt		requirements.txt
tc_merge_annotations.py		tc_merge_annotations.py
tc_reannotate.py		tc_reannotate.py
tc_update_gff3.py		tc_update_gff3.py
tc_utils.py		tc_utils.py
version.py		version.py
workflow_scheme.svg		workflow_scheme.svg

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

TideCluster

Installation

Usage

TideHunter step

Clustering step

Annotation step

TAREAN step

Run all steps

Example for full pipeline

Output

Tidehunter Step

Clustering Step

Annotation Step

TAREAN Step

Updating gff3 file based on manual annotation

Reannotation of tandem repeats using similarity based approach

Usage:

Credits

Authors

License

About

Releases 16

Packages

Languages

License

kavonrtep/TideCluster

Folders and files

Latest commit

History

Repository files navigation

TideCluster

Installation

Usage

TideHunter step

Clustering step

Annotation step

TAREAN step

Run all steps

Example for full pipeline

Output

Tidehunter Step

Clustering Step

Annotation Step

TAREAN Step

Updating gff3 file based on manual annotation

Reannotation of tandem repeats using similarity based approach

Usage:

Credits

Authors

License

About

Resources

License

Stars

Watchers

Forks

Releases 16

Packages 0

Languages

Packages