sTELLeR

sTELLeR; (s) Transposable ELements Long (e) Read

a tool for detection of non-reference transposable element using long-read data such as PacBio or ONT

sTELLeR is a fast and CPU light tool designed and tested for detection of transposable element insertion in long-read data, however it supports identification of any type of insertions. sTELLeR gives output in VCF, is haplotype-aware and can run on genome assemblies as well as on any species

Example scripts, bam test file and TE fasta sequences are available at https://github.com/kristinebilgrav/sTELLeR_supplementary/

Install

Option 1:

git clone https://github.com/kristinebilgrav/sTELLeR.git

Install dependencies:

• minimap2
• pysam
• samtools

All available through bioconda.

Option 2:

Dowload docker container containing sTELLeR available at: https://hub.docker.com/r/kristinebilgrav/steller

run using docker or singularity with:

python /sTELLeR/steller/steller.py 

Run

sTELLeR needs a bam or cram file, the reference used to align it, and a fasta file with insertions sequences to look for (--TE_fasta). This allows for detection of any type of insertions, not only TE insertions. Example TE sequences are given in the test dataset HERE
If the input is phased, the output VCF will contain information regarding which haplotype(s) the TE is on and its genotype. If not phased the haplotype will be 0 and the genotype will indicate if TE is heterozygous (1/0) or homozygous (1/1).

require python3, pysam, samtools, minimap2

python steller/steller.py [-h] [-R REF] -tf TE_FASTA -b BAM [-tr TE_REF] -s STYLE [-r SR] [-m MQ] [-k] [-o OUTPUT]

arguments:

-R REF, --ref     reference genome
-tf, --TE_fasta   fasta file with sequence of elements to be detected
                      
-b, --bam         bam file
                     
-s , --style      ont or pb (clr) / hifi

optional:

-k, --keep_intermediates Keep intermediate files
                    
-o, --output      Output file name

-r , --sr         Minimum number of supporting split reads/insertions to call a variant (default 3)

-m , --mq         Mapping quality (default 20)
                    
-mr, --maxreads   Maximum number of supporting split reads/insertions to call a variant (default 100)

The intermediate files (--keep_intermediates) include per-chromosome fasta files with candidate insertions as well as SAM files with aligned insertions.
Supporting reads (--sr) of a insertion depends on the coverage of the sample and the desired sensitivity (dafault=3).
Mapping quality (--mq) is a threshold of accepted TE calls to the specific TE sequence (default=20).
Maximum number of reads (--maxreads) can be adjusted to avoid calls in regions exceeding given coverage (default=100).

NOTE: If using container, the executable is:

python /sTELLeR/steller/steller.py 

Examples

Example to run on PB data:

python steller/steller.py --ref < genome ref_file > --TE_fasta < TE sequence fasta file > --bam < bamfile > --sr 4 --style pb -o < output prefix > -mr 80

Example to run on ONT data:

python steller/steller.py --ref < genome ref_file > --TE_fasta < TE sequence fasta file > --bam < bamfile > --sr 4 --style ont -o < output prefix > -mr 80

Example using test data located in: https://github.com/kristinebilgrav/sTELLeR_supplementary/

Download all files in repository and run:

python steller/steller.py --ref HG38_chr22.fa --TE_fasta fasta/TEfastasequences.fa --bam testdata.bam --sr 4 --style pb -o testdata_res -mr 80

Results will be given in testdata_res_repeats.vcf and correct results can be verified by comparing to the provided file testbamTRUTH.txt - a total of 7 variants should be identified.

Output

The output is given in a VCF file, which is compatible to merge with VCF files from other variant callers.
The variant is named INS:ME and in the info section the TE type along with start and end positions are given.
In the format section (TE:TS:HT:GT) items such as TE type (as given in the fasta file), TS transposon start position, HT haplotag (0 if not phased) and GT genotype (1/0 or 1/1 for unphased and 1|0, 0|1 or 1|1 for phased) are written out.

If not specified otherwise, intermediate files containing fasta sequences of candidate insertions (ID_chr_reads.fasta) and candiate insertions aligned to TEs (ID_chr_repeats.sam) will be removed.