How to detect de novo variants with trio-data

Question

[Shin Taguchi]

Hi, all.

I have a trio-dataset of a kind of fish with whole genome sequence: father, mother and offspring. I want to get information of de novo variant sites from this trio-dataset.

How can I check for variants present in descendants that are not present in either parent, and look for variant sites with Mendelian violations?

I'd appreciate any ideas on creating a methodology to go about this analysis.

Answer 1

[pragnapcu]

Look for KGGSEQ

[Shin Taguchi]

Thank you for your advice!

Answer 2

[Pierre Lindenbaum]

gatk: gatk VariantAnnotator -R "ref.fasta" -V fam.vcf -A PossibleDeNovo -O fam.denovo.vcf -ped fam.ped

bcftools https://samtools.github.io/bcftools/howtos/plugin.trio-dnm2.html

[Shin Taguchi]

@Pierre Lindenbaum

Thank you for your comment.

When using GATK, is the file specified in option -V a VCF file with the trio (father, mother, child) merged and variants called?

[Pierre Lindenbaum]

merged and variants called? yes

[Shin Taguchi]

Thank you very much!

I'm considering the following pipeline in broad terms. How do you think about this?

Raw data

Adapter trim (Trimmomatic)

Mapping (BWA)

Sorting and removing duplicates (samtools and picard)

Gnenotyping each individual (GATK: HaplotypeCaller)

Merging trio (GATK: CombineGVCFs)

Joint genotyping (GATK: GenotypGVCFs)

Select variants (GATK: SelectVariants)

Hard filtering (GATK: VariantFiltration)

Selecting de novo variants (GATK: VariantAnnotator)