Hi @FraCoppo
Thanks for sharing the history! Very helpful.
The error message is reporting that the input paired end reads did not have intact pairs.
Screenshot
We have a few topics about the issue. Searching with the error message can find them.
It looks like you used Trimmomatic for the QC. This tool will output intact pairs, but only if you input the reads from the same exact run. Your appear to be from mixed runs, even if they are putatively the same sample. You might be able to filter after if that was on purpose with a tool like Seqtk.
Or, you could switch to using a QA tool that will perform the clip and the filter all together. Cutadapt and fastp are two examples. Using collections can also help to avoid mixing up samples.
We have resources for this, and a simple workflow you could adapt to do everything in a batch. Change the trimming options in the Cutadapt tool to fit your data, or even swap out the trimming tool to use fastp instead if you want.
Hope this helps but let us know if you have any follow up questions! 