One row per each variant for each sample. We expand the FORMAT object into multiple rows (while duplicating all other information).
{
# all other values the same
"FORMAT" : {
"AF": 0.2 # AF for SAM1
},
"sample":"SAM1"
}
{
# all other values the same
"FORMAT" : {
"AF": 0.1 # AF for SAM2
},
"sample":"SAM2"
}
Annotation centric (previously):
One row per each annotation per each variant for each sample. We expand further expand the ANN object into multiple rows (while duplicating all other information).
{
# all other values the same
"INFO" : {
"ANN" : {
"effect": "one" # first ANN annotation for SAM1
}
},
"FORMAT" : {
"AF": 0.2
},
"sample":"SAM1"
}
{
# all other values the same
"INFO" : {
"ANN" : {
"effect": "two" # second ANN annotation for SAM1
}
},
"FORMAT" : {
"AF": 0.2
},
"sample":"SAM1"
}
{
# all other values the same
"INFO" : {
"ANN" : {
"effect": "one" # first ANN annotation for SAM2
}
},
"FORMAT" : {
"AF": 0.1
},
"sample":"SAM2"
}
{
# all other values the same
"INFO" : {
"ANN" : {
"effect": "two" # second ANN annotation for SAM2
}
},
"FORMAT" : {
"AF": 0.1
},
"sample":"SAM2"
}
Need different modes for outputting data in different orientations from the vcf atomizer:
Variant centric:
One row per unique variant. Multi-sample vcfs would have multiple sample objects under the FORMAT object.
Sample variant centric (current):
One row per each variant for each sample. We expand the FORMAT object into multiple rows (while duplicating all other information).
Annotation centric (previously):
One row per each annotation per each variant for each sample. We expand further expand the ANN object into multiple rows (while duplicating all other information).