The read cleaning (such as adapter remove, low quality base cut and polyA/T trimming) is required for FusionSeq? When we cleaned FASTQ file "SRR018259" which is publicly available, the FusionSeq cannot find correct fusions. On the other hand, when we run the FusionSeq without cleaning, it can find same fusions report in the paper. I would like to know which FASTQ (raw or cleaned) is recommended to load the FusionSeq.
And if cleaning is recommended, which is better to remove adapters.
(1) trimming adapters from reads (Read length is differ)
(2) removing the read itself (Read length is same)
thank you for interest in FusionSeq and for your note. We typically run
the programs on clean reads (i.e. without adapters), but we keep all
reads, regardless of the score. The filtering of potential artifactual
fusion transcripts is performed in subsequent steps within FusionSeq.
Please also note that the filtering step may require some tuning
depending on the specific library preparation protocol. Hence, I would
recommend to remove the reads that have adapters (no trimming).