Autosomal BNDs with GT "./." called

[dnil]

Unsure why we would set that - afik even if we can't quite pin the zygosity, they should at least be inferred to be on one allele, say "./1".

At a glance this seems to be for small BNDs only.

[dnil]

E.g.

7       37995773        SV_208_1        N       ]7:37995959]N   50      PASS    Annotation=SFRP4;CSQ=BND|intron_variant|MODIFIER|SFRP4|ENSG00000106483|Transcript|ENST00000447200|protein_coding||1/5||||||||||-1|cds_end_NF|HGNC|10778|||||ENSP00000402262||C9JMJ2_HUMAN|UPI000198CDF4|||Ensembl|||||||,BND|downstream_gene_variant|MODIFIER|EPDR1|ENSG00000086289|Transcript|ENST00000199448|protein_coding|||||||||||4231|1||HGNC|17572|YES|||CCDS5454.2|ENSP00000199448|EPDR1_HUMAN|D6RIH7_HUMAN|UPI000012EF12|||Ensembl|||||||;CTG=TAAATGTTTTCTCCTCCTCCCCTACATCCTTCTTTTTTCTTCTTCCATTCCTCTTCCTTCTCTTTCTTCTTCTCTTCTTCCTCTTCTTTTTCTCTTCTCCTTCTCCCCTCCCCCGCCACCTCCTCCTTCTCCACCTCCTCCTCCTCCACCTCCTCCTCCTGCTTCTCCTCCTCCCTCTTCCTCTTCCTCCTCCTCCTCCTCTTATTCTTCTTCCTCTTCCTCTTCTT;FOUNDBY=1;GeneticModels=positivefish:AD|AR_hom;LFA=0,12;LFB=0,12;LTE=0,2;REGIONA=37995757,37995827;REGIONB=37995846,37995985;RankResult=2|0|1|0|3|0|0|0;RankScore=positivefish:6.0;RankScoreMinMax=positivefish:-17.0:44.0;RankScoreNormalized=positivefish:0.3770491803278688;SUPP_VEC=001;SVTYPE=BND;most_severe_consequence=10778:BND|intron_variant,17572:BND|downstream_gene_variant;most_severe_pli=0.02;positivefish_tiddit_ACC20554A57_CHROM=SV_208_1|7,SV_208_2|7;positivefish_tiddit_ACC20554A57_FILTERS=SV_208_1|PASS,SV_208_2|PASS;positivefish_tiddit_ACC20554A57_INFO=SV_208_1|SVTYPE:BND|REGIONA:37995757:37995827|REGIONB:37995846:37995985|LFA:0:12|LFB:0:12|LTE:0:2|CTG:TAAATGTTTTCTCCTCCTCCCCTACATCCTTCTTTTTTCTTCTTCCATTCCTCTTCCTTCTCTTTCTTCTTCTCTTCTTCCTCTTCTTTTTCTCTTCTCCTTCTCCCCTCCCCCGCCACCTCCTCCTTCTCCACCTCCTCCTCCTCCACCTCCTCCTCCTGCTTCTCCTCCTCCCTCTTCCTCTTCCTCCTCCTCCTCCTCTTATTCTTCTTCCTCTTCCTCTTCTT,SV_208_2|SVTYPE:BND|REGIONA:37995757:37995827|REGIONB:37995846:37995985|LFA:0:12|LFB:0:12|LTE:0:2|CTG:TAAATGTTTTCTCCTCCTCCCCTACATCCTTCTTTTTTCTTCTTCCATTCCTCTTCCTTCTCTTTCTTCTTCTCTTCTTCCTCTTCTTTTTCTCTTCTCCTTCTCCCCTCCCCCGCCACCTCCTCCTTCTCCACCTCCTCCTCCTCCACCTCCTCCTCCTGCTTCTCCTCCTCCCTCTTCCTCTTCCTCCTCCTCCTCCTCTTATTCTTCTTCCTCTTCCTCTTCTT;positivefish_tiddit_ACC20554A57_POS=SV_208_1|37995773,SV_208_2|37995959;positivefish_tiddit_ACC20554A57_QUAL=SV_208_1|50,SV_208_2|50;positivefish_tiddit_ACC20554A57_SAMPLE=SV_208_1|ACC20554A57|GT:./.|CN:2|COV:27:26:31|DV:0|RV:2|LQ:0.0:0.0|RR:21:26|DR:28:40,SV_208_2|ACC20554A57|GT:./.|CN:2|COV:27:26:31|DV:0|RV:2|LQ:0.0:0.0|RR:21:26|DR:28:40;set=tiddit;svdb_origin=tiddit;swegenAF=0.001;swegenOCC=1;tiddit_CHROM=SV_208_1|7;tiddit_FILTERS=SV_208_1|PASS;tiddit_INFO=SV_208_1|SVTYPE:BND|REGIONA:37995757:37995827|REGIONB:37995846:37995985|LFA:0:12|LFB:0:12|LTE:0:2|CTG:TAAATGTTTTCTCCTCCTCCCCTACATCCTTCTTTTTTCTTCTTCCATTCCTCTTCCTTCTCTTTCTTCTTCTCTTCTTCCTCTTCTTTTTCTCTTCTCCTTCTCCCCTCCCCCGCCACCTCCTCCTTCTCCACCTCCTCCTCCTCCACCTCCTCCTCCTGCTTCTCCTCCTCCCTCTTCCTCTTCCTCCTCCTCCTCCTCTTATTCTTCTTCCTCTTCCTCTTCTT|SUPP_VEC:1;tiddit_POS=SV_208_1|37995773;tiddit_QUAL=SV_208_1|50;tiddit_SAMPLE=SV_208_1    GT:CN:COV:DV:RV:LQ:RR:DR        ./.:2:27,26,31:0:2:0,0:21,26:28,40

[dnil]

I guess it's these two fallbacks

TIDDIT/tiddit/tiddit_variant.pyx

Line 391 in cd3a786

GT="./."

that trigger with weird small BNDs that dont have good coverage info.

[J35P312]

Hello!
I will have a look!

This variant is called through de novo assembly, but is supported by two split reads. I think it shows that there is a need for better genotyping from the de novo asemblies, and perhaps better filtering, the contig looks really repetitive and messy ;).

I can do a quick fix and make so that the GT is updated to ./1 if it the variant is called by de novo assembly and there is no other support. Does it sound good?
Later on we could set the GT based on the number of reads that support the contig.

[dnil]

Thank you! Yes, that should cover more at least - there is usually just a handful per case anyway - and we can iterate if there are other types. I guess the key thing here really is keeping track of which individual had the tentative signal, and should go as GT "./1" or something similar, in contrast to ungenotyped individuals without signal, that would be fine to keep as "./.".