I did a pre-filtering for the sequences. Then I use a subset of a few thousand random good quality sequences to perform some test mappings.
When running as "unpaired" sequences (only -best is used, the rest is default), bowie can map about 80% of the total. A closer look at the output file shows about 70% (of the total) belong to the same pair. And about 30% (of the total) were mapped to the same chromosome and with the right distance and direction.
When running as "paired-end" sequences (-best is used with a large -X of a few kb, all others are default.), only 5% can be mapped and many of them showing different location (different chromosome and/or base positions when compared with the unpaired mapping result).
Can anyone help to see where things goes wrong here?
Thanks!
Nan
When running as "unpaired" sequences (only -best is used, the rest is default), bowie can map about 80% of the total. A closer look at the output file shows about 70% (of the total) belong to the same pair. And about 30% (of the total) were mapped to the same chromosome and with the right distance and direction.
When running as "paired-end" sequences (-best is used with a large -X of a few kb, all others are default.), only 5% can be mapped and many of them showing different location (different chromosome and/or base positions when compared with the unpaired mapping result).
Can anyone help to see where things goes wrong here?
Thanks!
Nan