VAF distribution of variants with PASS status

[mkazanov]

We plotted the VAF distribution for variants (obtained in Tumor-only mode) with PASS status, and it appears to be germline-like, with peaks at ~50% and ~100%.
What could be the reason for this?

[kishwarshafin]

Can you provide a little more details on what exactly you are running? One shot in the dark is that you likely are using normal from a different sample than the tumor.

[mkazanov]

We did not use a normal sample.
The analysis was run in tumor-only mode.

[kishwarshafin]

can you provide the command and some details on the sample?

[mkazanov]

I will able to provide the exact command tomorrow. Thank you for your help.

[mkazanov]

extra_args="--min_base_quality=1,
--min_mapping_quality=0,
vsc_min_count_snps=2,
vsc_min_count_indels=2,
vsc_min_fraction_snps=0.045,
vsc_min_fraction_indels=0.045,
--max_reads_per_partition=0,
--realign_reads=false"

docker run -u $(id -u):$(id -g) -v "/tmp:/tmp" -v "/mnt:/mnt" google/deepsomatic:1.9.0 run_deepsomatic --model_type=WGS_TUMOR_ONLY --ref="$ref" --reads_tumor="${LOCAL_BAM}" --output_vcf="${VCF}" \
      --sample_name_tumor="${SAMPLE}" --num_shards=10 --logging_dir="${DEEPSOMATIC_OUTPUT_DIR}/logs" --use_default_pon_filtering=true --make_examples_extra_args="${extra_args}" \
      --intermediate_results_dir $DEEPSOMATIC_TMP_DIR

[kishwarshafin]

@mkazanov in your output you should have a report similar to this:

https://storage.googleapis.com/deepvariant/deepsomatic-case-study-outputs/1.9.0/wgs-tumor-only/deepsomatic.output.visual_report.html

Can you paste it here? The report is pasted is WGS_TUMOR_ONLY for HCC1395 sample that we benchmark on.