how to avoid to getting repeat region ?

[SeqCrafter]

Hi Inigo,

I know that you use get_mate_intervals to build realign intervals, and you will use all the soft-clipped/hard-clipped reads and discordant reads in this function. But we know that if some eccDNA reads were from repeat regions (actually it's common), the aligner such as bwa will have multi-align reads including supplement reads and discordant reads mostly similar to eccDNA reads. Is it right?

So how can we avoid getting repeat regions in eccDNA? Will it work if I use --only_discordant?

Best wishes. Pan xiaoguang

[iprada]

Hi,

Read alignments to repeats and multi mapped regions should be handled by the probabilistic model. Reads with low probabilities are filtered out by default.

best,

Inigo