Estimating Contamination - Targeted Panel
205 views
Skip to first unread message
Robert Sicko
Jan 30, 2018, 4:03:15 PM1/30/18
to verifyBamID
Hi,
We are trying to detect sample contamination in batches of targeted sequencing samples (15 per run). No comparison genotype data available, but we want to detect if any samples contain a mixture of more than one sample. We created a mock contamination sample by spiking DNA from sample2 into sample1.
I ran verifyBamID (1.1.3) on each sample as follows:
./verifyBamID --vcf 180124-1000G_phase1.snps.high_confidence.hg19.intersected_w_scid.vcf --bam $sample_bam --out ${sample_bam}_verifybam --maxDepth 1000 --precise --ignoreRG
Results are shown below. I notice FREEMIX is >0.03 for all samples but as stated in the docs the contam sample does have a large FREELK1-FREELK0. I just don't know why FREEMIX is so large.
Thanks,
Bob
| #SEQ_ID | RG | CHIP_ID | #SNPS | #READS | AVG_DP | FREEMIX | FREELK1 | FREELK0 | FREE_RH | FREE_RA | CHIPMIX | CHIPLK1 | CHIPLK0 | CHIP_RH | CHIP_RA | DPREF | RDPHET | RDPALT | LK1-LK0 |
| NA12878 | ALL | NA | 190 | 105684 | 556.23 | 0.41738 | 12869.78 | 13016.96 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -147.18 |
| contam | ALL | NA | 190 | 99082 | 521.48 | 0.11943 | 13757.01 | 16077.28 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -2320.27 |
| 101 | ALL | NA | 190 | 122038 | 642.31 | 0.1473 | 18395.67 | 18793.14 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -397.47 |
| 102 | ALL | NA | 190 | 128412 | 675.85 | 0.1189 | 14683.43 | 14782.45 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -99.02 |
| 103 | ALL | NA | 190 | 67040 | 352.84 | 0.20848 | 6258.31 | 7008.78 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -750.47 |
| 104 | ALL | NA | 190 | 125581 | 660.95 | 0.22077 | 14120.24 | 14687 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -566.76 |
| 105 | ALL | NA | 190 | 48454 | 255.02 | 0.2325 | 4889.09 | 5294.43 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -405.34 |
| 106 | ALL | NA | 190 | 105552 | 555.54 | 0.03397 | 12108.93 | 12490.17 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -381.24 |
| 107 | ALL | NA | 190 | 95481 | 502.53 | 0.19454 | 9507.44 | 10291.24 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -783.8 |
| 107-2 | ALL | NA | 190 | 100201 | 527.37 | 0.11967 | 10317.52 | 10649.87 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -332.35 |
| 108 | ALL | NA | 190 | 105646 | 556.03 | 0.40653 | 11863.23 | 11966.78 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -103.55 |
| 109 | ALL | NA | 190 | 107761 | 567.16 | 0.21798 | 12910.33 | 13361.75 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -451.42 |
| 110 | ALL | NA | 190 | 2727 | 14.35 | 0.24626 | 436.96 | 480.72 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -43.76 |
| 111 | ALL | NA | 190 | 118092 | 621.54 | 0.40362 | 11095.55 | 11545.68 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -450.13 |
| 112 | ALL | NA | 190 | 59090 | 311 | 0.13762 | 8884.5 | 8924.07 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | -39.57 |
Hyun Min Kang
Feb 4, 2018, 11:55:34 PM2/4/18
to verif...@googlegroups.com
Bob, using only 190 SNPs may be too small to reliably estimate the contamination. Any chance that you could include more variants?
--
You received this message because you are subscribed to the Google Groups "verifyBamID" group.
To unsubscribe from this group and stop receiving emails from it, send an email to verifybamid...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Robert Sicko
Feb 9, 2018, 1:19:28 PM2/9/18
to verifyBamID
Our panel only targets 39 genes and each sample only contains ~300 variants after filtering. I could other reference VCFs if you think getting closer to 300 will make a difference (for the above example I interestected 1000G_phase1.snps.high_confidence with out ROI).
Thanks,
Bob
Hyun Min Kang
Feb 9, 2018, 6:16:40 PM2/9/18
to verif...@googlegroups.com
When estimating FREEMIX, you can still use all 1000 genomes variants within the target region, to increase the marker density, perhaps?
Hyun.
Reply all
Reply to author
Forward
0 new messages