Circular contig in multicontig reference remapping (SMRTPipe, BLASR)

[Gestahl]

Hello everybody,

I am doing a secondary analysis involving temporal data (IPDs) using SMRTPipe. I am using P_Mapping (BLASR) to map reads to a multicontig reference sequence. My reference genome consists of one circular (mt) and some linear reference sequences. This yields in low coverage (compared to the average coverage) at the "ends" of the circular reference sequence.

Please take a look at the coverage plot of my remapped mitochondrion in the attachment.

This is bad. I want to get rid of the error introduced due to reads not aligning to the artificial ends of the circular reference made linear or aligning with low quality somewhere else.

I want BLASR to allow overhangs, or even better: to treat my circular reference as circular.

- IS ONE OF THESE OPTIONS POSSIBLE?
- CAN IT BE DONE AND HOW?

Thank you for reading and thinking!

[rhall]

Blasr aligns all the way to the end of the sequence by default. From the plot you show it appear the reference has not been circularized correctly, i.e. the ends overlap and have not been trimmed. https://github.com/PacificBioscience...g-and-trimming