Skip to content

FAQ

Jump to bottom
Joe Zhu edited this page May 13, 2017 · 7 revisions

Frequently asked questions

Data filtering

Data filtering is an important step for deconvolution.

utilities/dataExplore.r -vcf data/exampleData/PG0415-C.eg.vcf.gz \
    -plaf data/exampleData/labStrains.eg.PLAF.txt \
    -o PG0415-C

PG0415_data

We observe a small number of heterozygous sites with high coverage (marked as crosses above), which can potentially mislead our model to over-fit the data with additional strains.

./dEploid -vcf data/exampleData/PG0415-C.eg.vcf.gz \
    -plaf data/exampleData/labStrains.eg.PLAF.txt \
    -noPanel -o PG0415-CNopanel -seed 2

utilities/interpretDEploid.r -vcf data/exampleData/PG0415-C.eg.vcf.gz \
    -plaf data/exampleData/labStrains.eg.PLAF.txt \
    -dEprefix PG0415-CNopanel \
    -o PG0415-CNopanel

PG0415_noFilter

The data exploration utility utilities/dataExplore.r identifies a list of potential outliers. After filtering,

./dEploid -vcf data/exampleData/PG0415-C.eg.vcf.gz \
    -plaf data/exampleData/labStrains.eg.PLAF.txt \
    -noPanel -o PG0415-CNopanel.filtered -seed 2 \
    -exclude PG0415-CPotentialOutliers.txt

utilities/interpretDEploid.r -vcf data/exampleData/PG0415-C.eg.vcf.gz \
    -plaf data/exampleData/labStrains.eg.PLAF.txt \
    -dEprefix PG0415-CNopanel.filtered \
    -o PG0415-CNopanel.filtered \
    -exclude PG0415-CPotentialOutliers.txt

PG0415_filtered

Clone this wiki locally