Process a list of TRED (trinucleotide repeats disease) loci, and infer the most likely genotype.

Make sure your Python version >= 2.7 (tested in ubuntu, Python 3 not yet supported):

pip install --user -U git+git://github.com/humanlongevity/tredparse.git

For accessing BAMs that are located on S3, please refer to docker/tredparse.dockerfile for installation of SAMTOOLS/pysam with S3 support.

Or, you can simply build and use the docker image:

cd docker
make build
docker run -v `pwd`:`pwd` -w `pwd` tanghaibao/tredparse \
    tred.py --tred HD test.bam

First specify the input bam paths and sample keys in a CSV file, like tests/samples.csv. This file is comma separated:

#SampleKey,BAM,TRED
t001,tests/t001.bam,HD
t002,tests/t002.bam,DM1

If third column is omitted, then all 30 TREDs are scanned. For example:

#SampleKey,BAM
t001,tests/t001.bam
t002,tests/t002.bam

Please also note that the BAM path can start with http:// or s3://, provided that the corresponding BAM index can be found.

Run tred.py on sample CSV file and generate TSV file with the genotype:

tred.py tests/samples.csv --workdir work

Highlight the potential risk individuals:

tredreport.py work/*.json --tsv work.tsv

The inferred "at-risk" individuals show up in results:

[DM1] - Myotonic dystrophy 1
rep=CAG inherit=AD cutoff=50 n=1 loc=chr19:45770205-45770264
SampleKey Calls DM1.FR                          DM1.PR     DM1.RR  DM1.PP
    t002  5/62   5|24  ...|1;39|1;40|1;42|1;43|1;46|2  49|3;50|8     1.0

[HD] - Huntington disease
rep=CAG inherit=AD cutoff=40 n=1 loc=chr4:3074877-3074933
SampleKey  Calls HD.FR                           HD.PR HD.RR  HD.PP
    t001  15/41  15|4  ...|1;21|1;24|2;29|1;34|1;41|1          1.0

One particular individual t001 appears to have 15/41 call (one allele at 15 CAGs and the other at 41 CAGs) at Huntington disease locus (HD). Since the risk cutoff is 40, we have inferred it to be at-risk.

A .report.txt file will also be generated that contains a summary of number of people affected by over-expanded TREDs as well as population allele frequency.

To better understand the uncertainties in the prediction, one call plot the likelihood surface based on the model. Using the same example as above at the Huntington disease case, we can run a command on the JSON output, with option --tred HD to specify the locus.

tredplot.py likelihood2 work/t001.json --tred HD

This generates the following plot: