Unusual MuTect behavior with --force_output option

mducarmducar Member
edited April 2015 in MuTect v1

Hi,

We are experiencing some strange results when we use the --force_output option with MuTect. We're running the latest version 1.1.7.

As an initial test, we compared the output from running MuTect as we usually do and dumpster diving for the rejects to running with the --force_output option. Example command lines are at the bottom.

When run in "normal" mode, a G>C mutation is considered but rejected -- rightfully so, there are only three C reads and all have low mapping quality, etc. However, the --force_output changes the variant to G>T. I looked at the bam file in IGV and there are zero "T" bases at this location. Mutect output is also included below.

Why is the --force_output giving these results?

Thanks,
Matt

# Regular run
/opt/java7/bin/java -Xmx4G -jar mutect-1.1.7.jar --analysis_type MuTect --reference_sequence Homo_sapiens_assembly19.fasta --cosmic Cosmic.vcf --dbsnp dbsnp.vcf --input_file:normal Normal.dedup.cleaned.bam --input_file:tumor Tumor.dedup.cleaned.bam --normal_panel PON_MuTect.vcf --coverage_file Tumor.coverage.wig --out Tumor.txt --intervals Check2Locus.interval_list --enable_extended_output

# Force output run
/opt/java7/bin/java -Xmx4G -jar mutect-1.1.7.jar --analysis_type MuTect --reference_sequence Homo_sapiens_assembly19.fasta --cosmic Cosmic.vcf --dbsnp dbsnp.vcf --input_file:normal Normal.dedup.cleaned.bam --input_file:tumor Tumor.dedup.cleaned.bam --normal_panel PON_MuTect.vcf --coverage_file Tumor.force.coverage.wig --out Tumor.forceout.txt --force_output --intervals Check2Locus.interval_list --enable_extended_output

# Regular run results
##MuTect:1.1.6-4-g69b7a37 Gatk:3.1-0-g72492bb
contig position context ref_allele alt_allele tumor_name normal_name score dbsnp_site covered power tumor_power normal_power normal_power_nsp normal_power_wsp total_reads map_Q0_reads init_t_lod t_lod_fstar t_lod_fstar_forward t_lod_fstar_reverse tumor_f contaminant_fraction contaminant_lod t_q20_count t_ref_count t_alt_count t_ref_sum t_alt_sum t_ref_max_mapq t_alt_max_mapq t_ins_count t_del_count normal_best_gt init_n_lod normal_f n_q20_count n_ref_count n_alt_count n_ref_sum n_alt_sum power_to_detect_positive_strand_artifact power_to_detect_negative_strand_artifact strand_bias_counts tumor_alt_fpir_median tumor_alt_fpir_mad tumor_alt_rpir_median tumor_alt_rpir_mad observed_in_normals_count failure_reasons judgement
22 29083962 TTCxACT G C Tumor Normal 0 COSMIC COVERED 0.999622 0.999622 1 1 1 196 5 -24.680031 0 0 0 0 0.02 1.842287 85 82 0 2879 0 60 0 0 0 GG 25.852911 0.018182 106 108 2 3742 63 0.010109 0.014489 (31,61,0,0) n/a n/a n/a n/a 0 fstar_tumor_lod,possible_contamination,alt_allele_in_normal,poor_mapping_region_alternate_allele_mapq REJECT

# Force output results
##MuTect:1.1.6-4-g69b7a37 Gatk:3.1-0-g72492bb
contig position context ref_allele alt_allele tumor_name normal_name score dbsnp_site covered power tumor_power normal_power normal_power_nsp normal_power_wsp total_reads map_Q0_reads init_t_lod t_lod_fstar t_lod_fstar_forward t_lod_fstar_reverse tumor_f contaminant_fraction contaminant_lod t_q20_count t_ref_count t_alt_count t_ref_sum t_alt_sum t_ref_max_mapq t_alt_max_mapq t_ins_count t_del_count normal_best_gt init_n_lod normal_f n_q20_count n_ref_count n_alt_count n_ref_sum n_alt_sum power_to_detect_positive_strand_artifact power_to_detect_negative_strand_artifact strand_bias_counts tumor_alt_fpir_median tumor_alt_fpir_mad tumor_alt_rpir_median tumor_alt_rpir_mad observed_in_normals_count failure_reasons judgement
22 29083962 TTCxACT G T Tumor Normal 0 COSMIC COVERED 0.999622 0.999622 1 1 1 196 5 -24.680031 0 0 0 0 0.02 0 85 82 0 2879 0 60 0 0 0 GG 32.504613 0 106 108 0 3742 0 0.010109 0.014489 (31,61,0,0) n/a n/a n/a n/a 0 fstar_tumor_lod,possible_contamination,poor_mapping_region_alternate_allele_mapq REJECT

Best Answer

Answers

  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie

    Hi Matt,

    Not sure what's going on here. Would you be able to share a snippet of data for us to test locally?

Sign In or Register to comment.