PostprocessGermlineCNVCalls generate a segment file with missing GT field for the segment file

Follow

NawarDalila

• November 09, 2022 13:07

Hi,

I am running the CNV workflow as recommended from this source:

https://github.com/broadinstitute/gatk/blob/master/scripts/cnv_wdl/germline/cnv_germline_cohort_workflow.wdl

I am using GATK v4.2.6.1 and I am getting the segment vcf file output as the foolowing:

 

#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	Patient1
chr1	2228507	CNV_chr1_2228507_237832907	G	.	3076.53	.	END=237832907	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:673:56:3077:137:129
chr2	21001470	CNV_chr2_21001470_223967294	C	.	3076.53	.	END=223967294	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:794:53:3077:56:112
chr3	4516232	CNV_chr3_4516232_186744045	T	.	3076.53	.	END=186744045	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:474:65:3077:99:147
chr4	52023697	CNV_chr4_52023697_186288874	T	.	3076.53	.	END=186288874	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:215:69:3077:146:100
chr5	218096	CNV_chr5_218096_177409787	G	.	3076.53	.	END=177409787	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:196:73:3077:135:113
chr6	6145359	CNV_chr6_6145359_160611981	T	.	3076.53	.	END=160611981	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:484:64:3077:103:144
chr6	160633493	CNV_chr6_160633493_160635564	G	<DUP>	15.84	.	END=160635564	GT:CN:NP:QA:QS:QSE:QSS	./.:5:2:14:16:18:16
chr6	160640407	CNV_chr6_160640407_160753321	C	.	1562.01	.	END=160753321	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:23:56:1562:89:56
chr7	1044286	CNV_chr7_1044286_151876880	T	.	3076.53	.	END=151876880	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:392:49:3077:94:126
chr8	11542965	CNV_chr8_11542965_143215875	T	.	3076.53	.	END=143215875	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:58:64:3077:119:153
chr9	116540	CNV_chr9_116540_136687617	T	.	3076.53	.	END=136687617	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:259:59:3077:117:98
chr10	18140477	CNV_chr10_18140477_119677542	T	.	3076.53	.	END=119677542	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:222:57:3077:108:125
chr11	532376	CNV_chr11_532376_128916991	G	.	3076.53	.	END=128916991	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:304:67:3077:86:108
chr12	2053303	CNV_chr12_2053303_124863980	G	.	3076.53	.	END=124863980	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:458:54:3077:92:91
chr13	36848416	CNV_chr13_36848416_113149777	A	.	2733.06	.	END=113149777	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:49:56:2733:95:106
chr14	23381780	CNV_chr14_23381780_94383497	T	.	3076.53	.	END=94383497	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:245:63:3077:102:118
chr15	34788279	CNV_chr15_34788279_99713031	G	.	3076.53	.	END=99713031	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:280:65:3077:104:136
chr16	854151	CNV_chr16_854151_86513345	A	.	3076.53	.	END=86513345	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:147:62:3077:135:104
chr17	1744736	CNV_chr17_1744736_80119591	A	.	3076.53	.	END=80119591	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:215:58:3077:143:112
chr18	3067002	CNV_chr18_3067002_59359504	T	.	3076.53	.	END=59359504	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:226:60:3077:130:132
chr19	4090338	CNV_chr19_4090338_55157849	T	.	3076.53	.	END=55157849	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:310:49:3077:84:126
chr20	6769867	CNV_chr20_6769867_63498023	C	.	3076.53	.	END=63498023	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:81:71:3077:123:123
chr21	34370207	CNV_chr21_34370207_43072453	T	.	1040.82	.	END=43072453	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:17:73:1041:121:100
chr22	19876845	CNV_chr22_19876845_50524671	C	.	3076.53	.	END=50524671	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:78:47:3077:99:142
chrX	31121573	CNV_chrX_31121573_154863492	T	.	3076.53	.	END=154863492	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:357:56:3077:51:134
chrX	154885874	CNV_chrX_154885874_154886417	T	<DEL>	50.99	.	END=154886417	GT:CN:NP:QA:QS:QSE:QSS	0/1:1:1:51:51:51:51
chrX	154895817	CNV_chrX_154895817_155022812	A	.	1549.89	.	END=155022812	GT:CN:NP:QA:QS:QSE:QSS	0/0:2:25:83:1550:112:51

 

First, I noticed that this is NOT the usual VCF output we used to get from older GATK versions where 0 is NORMAL ploidy, 1 is DELETION, and 2 is DUPLICATION.

Second, I noticed that for this results (and other samples) we have (./.) for the GT field for the duplication on chr6 while it is 0/1 for the deletion in chrX. I checked other samples and it is always the duplication that leads to a missing GT field, is that normal?

Finally I checked the INTERVAL VCF file generated by the same tool for the same sample and I didn't find the same problem as you can see below for the two regions mentioned in chr6 and chrX:

chr6	160606217	CNV_chr6_160606217_160606918	N	<DEL>,<DUP>	.	.	END=160606918	GT:CN:CNLP:CNQ	0:2:100,100,0,100,100,100:100
chr6	160611302	CNV_chr6_160611302_160611981	N	<DEL>,<DUP>	.	.	END=160611981	GT:CN:CNLP:CNQ	0:2:100,100,0,87,100,100:87
chr6	160633493	CNV_chr6_160633493_160634172	N	<DEL>,<DUP>	.	.	END=160634172	GT:CN:CNLP:CNQ	2:5:99,99,72,65,15,0:15
chr6	160634863	CNV_chr6_160634863_160635564	N	<DEL>,<DUP>	.	.	END=160635564	GT:CN:CNLP:CNQ	2:5:99,99,72,43,14,0:14
chr6	160640407	CNV_chr6_160640407_160641108	N	<DEL>,<DUP>	.	.	END=160641108	GT:CN:CNLP:CNQ	0:2:100,58,0,67,60,66:58
chr6	160645954	CNV_chr6_160645954_160646655	N	<DEL>,<DUP>	.	.	END=160646655	GT:CN:CNLP:CNQ	0:2:100,89,0,90,100,100:89
chr6	160650078	CNV_chr6_160650078_160650757	N	<DEL>,<DUP>	.	.	END=160650757	GT:CN:CNLP:CNQ	0:2:100,100,0,100,100,100:100

 

chrX	154862823	CNV_chrX_154862823_154863492	N	<DEL>,<DUP>	.	.	END=154863492	GT:CN:CNLP:CNQ	0:2:100,100,0,100,100,100:100
chrX	154885874	CNV_chrX_154885874_154886417	N	<DEL>,<DUP>	.	.	END=154886417	GT:CN:CNLP:CNQ	1:1:99,0,51,100,100,100:51
chrX	154895817	CNV_chrX_154895817_154896492	N	<DEL>,<DUP>	.	.	END=154896492	GT:CN:CNLP:CNQ	0:2:100,100,0,100,100,100:100

 

Would anybody explain to me this difference in the vcf format between the interval and segment files please? And is it a change that was expected from the old version to the new one?

Best regards

Nawar

0

• 

  Anthony DiCi

  • November 22, 2022 21:26

  Hi NawarDalila,

  Thank you for writing to the GATK forum! I hope that we can help you sort this out.

  The duplications are no-call on purpose because we can't phase them with the tools that we have. So, we don't know if the duplication is one copy on one haplotype and three copies on the other or two copies on each haplotype, which is the information we need to know if we gave it a genotype. So, we don't give it a genotype. Which we recognize is different from SNPs and Indels, which can make some things more complicated, but this is the correct way to represent it.

  Please find more detailed information on using PostprocessGermlineCNVCalls here.

  I hope this helps to clear things up! Please do not hesitate to reach out if you have any other questions.

  Best,
  Anthony

  0

  Comment actions Permalink
• 

  Anthony DiCi

  • November 30, 2022 16:24

  Hi NawarDalila,

  We haven't heard from you in a while so we're going to close out this ticket. If you still require assistance, simply respond to this email and we'll be happy to pick up where we left off!

  Kind regards,

  Anthony​

  0

  Comment actions Permalink
• 

  Jesús María López Martí

  • December 07, 2022 13:46

  Hi Anthony Dias-Ciarla,

     regarding the same topic, I have some doubts about how to apply variant filters to DUP genotypes called as './.'. I have tried the tool VariantFiltration (GATK 4.3.0) using different parameters (isNoCall, isCalled,...) to filter those low quality DUP variants. Even I have tried to use inverted filters or the variantContext to keep only the not called genotypes in order to get de DUP variants but everything I have tried ended up ignoring all the DUP variants. On the other hand, the tool SelectVariant  works fine and I am able to keep/filter DUP variants using the same parameters that are not working for VariantFiltration. I have read all the online documentation and I didn't find anything at all. However, maybe I am missing something here. How I should proceed to filter some of the DUP genotypes.

  Best,

  Jesús M

   

   

   

  0

  Comment actions Permalink

Please sign in to leave a comment.