Hi, I have an analysis that stops the pipeline at MISO_RUN step because I have 3 samples with read sizes of 76pb and all the others are at 150pb. When I set

--miso_read_len to 150

, I get an error telling me that in the 76pb samples, miso finds a different read length than 150pb. I try to debug these samples directly manually by replacing

--read-len 150

with

--read-len 75

in the .command.sh files of the corresponding working folders and running them via bash .command.run but when I restart the pipeline, the samples debugged in this way are not cached. Does anyone have any idea how to fix this? Thanks in advance for your help.

If you want to compare the 76-bp samples to the 150-bp ones, the cleanest solution would be to trim the 150-bp reads to 76 bp. This is a shame, because especially for splicing analysis read length is helpful. Let me start with saying that MISO is not really designed for replicates. So if you have replicates, I honestly don't see the point of using it -- rMATS has the same function, but can deal with replicates and allows for variable read-length (you still have to specify a length, ideally some average, but with the parameter it will not crash the pipeline). If you don't have replicates or really want to use MISO and don't want to unnecessarily trim reads, you can just separate out the comparisons and run them at different length (for anything where you need to compare across lengths, you'll still have to trim). (Sorry I did not see this earlier.)