[simonandrews]

See this comment earlier in the thread. For most illumina libraries you can simply use the common adapter start sequence to trim and it will work for all variants.

[simonandrews]

Why can't you ask from the person who gave you the data?

We've actually now instigated a generic barcode mapping and adapter trimming analysis on all of our runs as the number of times the information we got from the submitting user didn't match up was causing us real problems. We now work out what we think is in the data from the sequence itself and then try to tie this to what the users submitted, not the other way around.

[Wallysb01]

We've actually now instigated a generic barcode mapping and adapter trimming analysis on all of our runs as the number of times the information we got from the submitting user didn't match up was causing us real problems. We now work out what we think is in the data from the sequence itself and then try to tie this to what the users submitted, not the other way around.

There is always just plain contamination too. In RNA-seq assemblies, I’ve pulled out heaps of fully assembled adapters that supposedly shouldn’t have been anywhere on the flow cell. And if you go to submit those assemblies to TSA, they will find them for you and you’ll have to remove them post-assembly. That is much harder than removing them from the reads...

So, I think its a good idea to trim based on every Illumina adapter, if you’re doing something that actually cares about removing any adapters at all.

[alexdobin]

If you map RNA-seq to a reference genome with STAR, each read will be automatically clipped as much as needed for alignment even if the adapter sequence is not known. You can later look at the soft-clipped sequences to figure out what the actual adapter sequence is.