search

Santy-8128 / Minimac3

Minimac3 is a low memory and computationally efficient implementation of the genotype imputation algorithms. Minimac3 is designed to handle very large reference panels in a more computationally efficient way with no loss of accuracy.
28 stars 12 forks source link

having issues converting freeze8 chromosome VCFs to the m3vcf format #17

Closed khzhu closed 1 year ago

khzhu commented 1 year ago

Hi, I have been following your online document and tried to create a TopMed reference panel. Freeze8 VCFs were downloaded from the public Bravo browser. I am having issues converting chromosome VCFs to the m3vcf format. The command I ran was as following: Minimac3 --refHaps chr3.BRAVO_TOPMed_Freeze_8.vcf.gz --processReference --prefix m3vcfs_test --mychromosome chr3 But I got

Loading Reference Haplotype Set from VCF File       : chr3.BRAVO_TOPMed_Freeze_8.vcf.gz

 Reading VCF File to calculate number of records ...

 Number of Markers read from VCF File                : 55503055
 Number of Markers with more than Two Alleles        : 0
 Number of Markers failing FILTER = PASS             : 6791670
 Number of Markers with inconsistent Ref/Alt Allele  : 0
 Number of Markers with duplicate ID:POS:REF:ALT     : 0
 Number of Insertions                                : 1688002
 Number of Deletions                                 : 3446271

 Number of Markers to be Recorded                    : 55503055
 Number of Haplotypes to be Recorded                 : 2

 Starting to load data ...

 Number of Threads to be Used = 1

 Empty Value for Individual : TOPMED at Marker : chr3:10009
 Most probably a corrupted VCF file. Please check input VCF file !!!
Aborted (core dumped)

as you can see there is a value for the position

chr3    10009   TOPMed_freeze_8?chr3:10,009     A       C       80      SVM;DISC;EXHET  VRT=1;NS=132345;AN=264690;AC=1380;AF=0.00521365;Het=1380;Hom=0  NA:FRQ  132345:0.00
khzhu commented 1 year ago

Freeze 8 VCF files downloaded from BRAVO are not standard multiple sample gVCFs, that caused the issue, not an issue with minimac3.