Example of using ScatTR

Running the example

In this example, we are estimating the copy number of a single TR in a simulated WGS sample. To run the example, make sure you have scattr or cargo installed in your PATH:

./run.sh

Input files

In the input/ directory, there are example input files to run ScatTR. In general, running ScatTR requires the following input files:

• A tab-delimeted catalog containing TR loci of interest: catalog.tsv
  • The file is expected to have columns: id, contig, start, end, and motif
• Aligned reads of the sample and their index: input/sample.bam and input/sample.bam.bai
• A reference genome and its index: input/reference.fa and input/reference.fa.fai

Output files

In the output/ directory, there are output files generated by ScatTR based on the example input files. ScatTR produces the following output files:

• The extracted bag of reads: sample.bag.bam
• The extracted depth and insert distributions: sample.stats.json
  • By default, ScatTR produces plots of these distributions (sample.depth_distr.png and sample.insert_distr.png). The --no-plot option disables this behavior
• The optimization problem definitions: sample.defs.json
  • This file information about all the reads associated with each TR and their relative positions with the decoy references (refer to manuscript methods section for more details)
• The estimated copy numbers: sample.genotypes.json
  • This file contains estimates for each TR locus in addition to a 95% confidence interval

Generating input files

Generating the input files requires having python, art_illumina, bwa, samtools in your PATH. The input files are generated by running:

./scripts/make_inputs.sh

The script creates a random reference genome (~200 kbp in length) that is used to generate reads for a heteroyzygous TR expansion with normal allele copy number being 3 and with expanded copy number being 100. The motif of the TR is CAGATA.