Each pair should be concatenated and separated by a N character.
The k-mers overlapping that position should have a specific marker in the output, e.g. C (concatenation).
Example
k=4
in1.fq:
@read1/1
ACGT
+
!!!!
...
in2.fq:
@read1/2
TGCA
+
!!!!
...
Extended Kraken format
output:
U read1 0 8 ref1:1 C:4 ref1:1
Bitmask output format (#14 )
The hit and coverage masks should not contain the concatenation k-mers, but the two reads should be separated by a pipe (|).
read1 ref1 8 2 8 1|1
Alternative solutions
If there is a cleaner way to obtain the same result without concatenating reads with N, we should consider it (e.g. query the two parts indipendently).
Add support for paired-end reads to the
querycommand.Updated specification
Behavior
Each pair should be concatenated and separated by a
Ncharacter. The k-mers overlapping that position should have a specific marker in the output, e.g.C(concatenation).Example
k=4
in1.fq:in2.fq:Extended Kraken format
output:
Bitmask output format (#14 )
The hit and coverage masks should not contain the concatenation k-mers, but the two reads should be separated by a pipe (
|).Alternative solutions
If there is a cleaner way to obtain the same result without concatenating reads with
N, we should consider it (e.g. query the two parts indipendently).