DANTE_LTR

• Citation
• Principle of DANTE_LTR
• Availability
• Installation
• Quick start guide - How to use DANTE and DANTE_LTR on Galaxy server
• Quick start guide - How to use command line version of DANTE and DANTE_LTR
• Tools description
• GFF3 DANTE_LTR output specification
• Modifying LTR-RT search constrains

Tool for identifying complete LTR retrotransposons based on analysis of protein domains identified with the DANTE tool. Both DANTE and DANTE_LTR are available on Galaxy server.

Citation

Novak, P., Hostakova, N., Neumann, P., Macas, J. (2024) – DANTE and DANTE_LTR: lineage-centric annotation pipelines for long terminal repeat retrotransposons in plant genomes. NAR Genomics and Bioinformatics 6:113. [https://doi.org/10.1093/nargab/lqae113]

Principle of DANTE_LTR

Complete retrotransposons are identified as clusters of protein domains recognized by the DANTE tool. The domains in the clusters must be assigned to a single retrotransposon lineage by DANTE. In addition, the orientation and order of the protein domains, as well as the distances between them, must conform to the characteristics of elements from REXdb database Neumann et al. (2019). In the next step, the 5' and 3' regions of the putative retrotransposon are examined for the presence of 5' and 3' long terminal repeats. If 5'- and 3'-long terminal repeats are detected, detection of target site duplication (TSD) and primer binding site (PSB) is performed. The detected LTR retrotranspsons are classified into 5 categories:

• Elements with protein domains, 5'LTR, 3'LTR, TSD and PBS - rank DLTP.
• Elements with protein domains, 5'LTR, 3'LTR, and PBS (TSD was not found) Rank DLP.
• Elements with protein domains, 5' LTR, 3'LTR, TSD (PBS was not found) - rank DLT.
• Elements with protein domains, 5'LTR and 3'LTR (PBS and TDS were not found) - rank DL.
• Elements as clusters of protein domains with the same classification, no LTRs - rank D.

Availability

DANTE_LTR and DANTE are available on Galaxy server or can be installed using conda package manager.

Installation:

conda create -n dante_ltr -c bioconda -c conda-forge -c petrnovak dante_ltr

Important version information DANTE_LTR versions up to 0.3.5.3 are compatible with REXdb Viridiplante database version 3.0. Versions >=4.0.1 are compatible with REXdb Viridiplante database version 3.0 and 4.0. REXdb Viridiplantae v 4.0 include additional LTR-RT lineages characterized non-angiosperm species. Updated REXdb and used classification system can be found https://github.com/repeatexplorer/rexdb.

Quick start guide - How to use DANTE and DANTE_LTR on Galaxy server

Detailed tutorial on how to use DANTE and DANTE_LTR on Galaxy server is here.

Quick start guide - How to use command line version of DANTE and DANTE_LTR

Installation of both DANTE and DANTE_LTR using conda into single environment:

conda create -n dante_ltr -c bioconda -c conda-forge -c petrnovak dante_ltr=0.4.0.1 dante=0.2.0
conda activate dante_ltr

Download example data:

wget https://raw.githubusercontent.com/kavonrtep/dante_ltr/main/test_data/sample_genome.fasta

Run DANTE on sample genome using 10 cpus:

dante -q sample_genome.fasta -o DANTE_output.gff3 -c 10

Output will contain annotation of individual protein domains identified by DANTE stored in GFF3 file named DANTE_output.gff3. Check DANTE documentation for more details (https://github.com/kavonrtep/dante)

Identify complete LTR retrotransposons from DANTE ouput using DANTE_LTR

dante_ltr -g DANTE_output.gff3 -s sample_genome.fasta -o DANTE_LTR_annotation -M 1

Option -M 1 will allow one missing domain in the complete LTR retrotransposon.

Output files will include:

• DANTE_LTR_annotation.gff3 - Annotation of all identified elements
• DANTE_LTR_annotation_statistics.csv - number of elements in individual categories
• DANTE_LTR_annotation_summary.html - graphical summary of the results

Create library of LTR RTs for similarity based annotation

dante_ltr_to_library -g DANTE_LTR_annotation.gff3 -s sample_genome.fasta -o LTR_RT_library.fasta

This step will create non-redundant library of LTR-RT sequences suitable for similarity based annotation using RepeatMasker.

Tools description

Detection of complete LTR retrotransposons

usage: dante_ltr [-h] -g GFF3 -s REFERENCE_SEQUENCE -o OUTPUT [-c CPU]
                 [-M MAX_MISSING_DOMAINS] [-L MIN_RELATIVE_LENGTH] [-S MAX_CHUNK_SIZE]
                 [-v] [--te_constrains TE_CONSTRAINS] [--no_ambiguous_domains]

        Tool for identifying complete LTR retrotransposons based on 
        analysis of protein domains identified with the DANTE tool
        

options:
  -h, --help            show this help message and exit
  -g GFF3, --gff3 GFF3  gff3 file with full output from Domain Based Annotation of Transposable Elements (DANTE)
  -s REFERENCE_SEQUENCE, --reference_sequence REFERENCE_SEQUENCE
                        reference sequence as fasta file
  -o OUTPUT, --output OUTPUT
                        output file path and prefix
  -c CPU, --cpu CPU     number of CPUs
  -M MAX_MISSING_DOMAINS, --max_missing_domains MAX_MISSING_DOMAINS
  -L MIN_RELATIVE_LENGTH, --min_relative_length MIN_RELATIVE_LENGTH
                        Minimum relative length of protein domain to be considered for retrostransposon detection
  -S MAX_CHUNK_SIZE, --max_chunk_size MAX_CHUNK_SIZE
                        
                                If size of reference sequence is greater than this value, reference is '
                                'analyzed in chunks of this size. default is 100000000 '
                                'Setting this value too small  will slow down the analysis
                                
  -v, --version         show program's version number and exit
  --te_constrains TE_CONSTRAINS
                        csv table specifying TE constraints for LTR search, template for this table 
                        can be found in https://github.com/kavonrtep/dante_ltr/blob/main/databases/lineage_domain_order.csv
  --no_ambiguous_domains
                        Remove ambiguous domains from analysis

Example:

mkdir -p tmp
./dante_ltr -g test_data/sample_DANTE.gff3 -s test_data/sample_genome.fasta -o tmp/ltr_annotation

Files in the output of extract_putative_ltr.R:

• prefix.gff3 - annotation of all identified elements
• prefix_D.fasta - partial elements with protein domains
• prefix_DL.fasta - elements with protein domains and LTR
• prefix_DLTP.fasta - elements with domains, LTR, TSD and PBS
• prefix_DLP.fasta - elements with domains, LTR and PBS
• prefix_DLT.fasta - elements with domains, LTR, TSD
• prefix_statistics.csv - number of elements in individual categories
• prefix_summary.html - graphical summary of the results

Making library of LTR RTs for RepeatMasker

If you want to annotate LTR RT elements with custom library using similarity based approach, you can use dante_ltr_to_library script wich will create non-redundant library which is formatted for RepeatMasker:

usage: dante_ltr_to_library [-h] -g GFF3 -s REFERENCE_SEQUENCE -o OUTPUT_DIR [-m MIN_COVERAGE] [-c CPU]

Creation of repeat library from dante_ltr output. Extract sequences based on gff3 inpute and reference fasta file. Run mmseqs2 clustering to cluster similar sequences to reduce library size. Exclude
clusters which have conflicting annotations and coverage below specified threshold.

options:
  -h, --help            show this help message and exit
  -g GFF3, --gff3 GFF3  gff3 file
  -s REFERENCE_SEQUENCE, --reference_sequence REFERENCE_SEQUENCE
                        fasta file
  -o OUTPUT_DIR, --output_dir OUTPUT_DIR
                        output directory
  -m MIN_COVERAGE, --min_coverage MIN_COVERAGE
                        Minimum coverage of cluster to be included in repeat library (default: 3)
  -c CPU, --cpu CPU     Number of cpus to use

GFF3 DANTE_LTR output specification

Types of features in GFF3:

• target_site_duplication: This feature represents sort direct repeats around the insertion site of a transposable element.
• transposable_element: This is the main feature representing the full extent of a transposable element within the genome.
• long_terminal_repeat (LTR): These are the direct repeats at the element's termini.
• protein_domain: This feature indicates a specific polyprotein domain identified by DANTE.
• primer_binding_site: This feature represents the site where a tRNA primer binds to initiate reverse transcription.

Attributes of features in GFF3:

• Rank: Rank of the elements (D, DL, DLT, DLP, DLTP) as described above
• Parent: Indicates the parent feature of the current feature, here a transposable element ID.
• Ndomains: The number of protein domains found within a transposable element.
• ID: A unique identifier for the feature.
• LTR_Identity: The percentage of sequence identity between 5' and 3' LTR sequences.
• LTR5_length and LTR3_length: The lengths of the 5' and 3' LTRs, respectively.
• TSD (Target Site Duplication): The sequence of the target site duplication.
• Final_Classification: A hierarchical classification of the transposable element based on REXdb classification system
• Name: The attibute is part of DANTE output and correspod to name of protein domain (RT. RH, PROT, ...)
• trna_id: The identifier for the tRNA related to the primer binding site.
• PBS_evalue: The E-value associated with the primer binding site.
• Best_Hit: Information about the best match of the protein domain to a known database entry.
• Best_Hit_DB_Pos: Position of the best hit within the database sequence.
• DB_Seq: The database sequence that corresponds to the best hit.
• Region_Seq: The sequence of the region where the query sequence was aligned to the database sequence.
• Query_Seq: The sequence of the query used to find the best hit.
• Identity: The percentage identity of the best hit match.
• Similarity: The similarity score of the best hit match.
• Relat_Length: The relative length of the match compared to the database sequence.
• Relat_Interruptions: Indicates the relative number of interruptions in the domain sequence.Interuption could be either stop codon or frameshift.
• Hit_to_DB_Length: The length of the hit compared to the database sequence length.

Modifying LTR-RT search constrains

It is possible to modify constraints for LTR search by providing a csv table with constraints for individual lineages.

The table has the following format:

Lineage	Domains order	offset5prime	offset3prime	domain_span	ltr_length
Class_I/LTR/Ty1_copia/Ale	GAG PROT INT RT RH	2000	2000	5700	123
Class_I/LTR/Ty1_copia/Alesia	GAG PROT INT RT RH	2000	3000	5400	273
Class_I/LTR/Ty1_copia/Angela	GAG PROT INT RT RH	6000	3000	5500	1074
Class_I/LTR/Ty1_copia/Bianca	GAG PROT INT RT RH	3500	3000	6000	132
Class_I/LTR/Ty1_copia/Bryco	GAG PROT INT RT RH	3000	3000	5000	287
Class_I/LTR/Ty1_copia/Gymco-I	GAG PROT INT RT RH	3500	2500	5400	151
Class_I/LTR/Ty1_copia/Gymco-II	GAG PROT INT RT RH	2000	6000	4600	156
Class_I/LTR/Ty1_copia/Gymco-III	GAG PROT INT RT RH	2000	2000	5400	247
Class_I/LTR/Ty1_copia/Gymco-IV	GAG PROT INT RT RH	2000	2000	5400	276
Class_I/LTR/Ty1_copia/Ikeros	GAG PROT INT RT RH	6500	3000	6100	359
...	...	...	...	...	...

• The Domain order column defines the order of individual protein domains required for positive detection of the elements.

• The number in the offset5prime column is the size of the upstream region used to search for the 5' LTR, while the offset3prime is the size of the downstream region used to search for the 3' LTR (These values correspond to yellowish and greenish boxes in the figure above).

• The domain_span is the maximal distance between N' end of first domain and the C' end of the last domain of the element.

Modify these constraints if you think that the default constraints lead to under-detection of elements whose structure deviates from the default constraints. Setting offset5prime, offset3prime or domain_span too high can however lead to the detection of aberrant or chimeric elements.

• The ltr_length is the shortest LTR for given lineage in REXdb database. The use of this value is currently not implemented in DANTE_LTR.

To use modified constrains use dante_ltr with option --te_constrains and provide the path to the modified csv table.

The full table with default constraints can be found in
databases/lineage_domain_order.csv.