These are great details! Thank you!
Suggestions
-
Consider running FastQC/Falco after the Cutadapt trimming step, then view all the reports together in MultiQC. This step makes sure that the trimming did what you expected. (actually remove adaptor/artifact and nothing else odd was detected). The reports are discussed in this tutorial with linkouts to the author’s site with more! → Hands-on: Quality Control / Quality Control / Sequence analysis
We have a public demonstration template workflow here that you can try! The input is just the original raw paired end collection of reads, the output are all the reports and trimmed reads. Be sure to allow it to process completely (data will show/hide throughout the runtime, and that’s expected!)
-
For mapping, we have some help for investigating the results closer in this tutorial. → Hands-on: Reference-based RNA-Seq data analysis / Reference-based RNA-Seq data analysis / Transcriptomics (mapping section!). Coverage graphs and zooming in on a few known genes in a genome browser can both be informative!
-
Finally, for the annotation choice, limiting to the basic genes on the primary chromosomes might help. This is the 4th option down on the Gencode website here → GENCODE - Human Release 49 and has in the description: This is the main annotation file for most users
We hope this feedback helps! There could be an issue with the reads of course! If this was from a public sample like SRA, you could review the metrics they capture and maybe there is discussion online at a site like Biostars.org. If these were your own samples, chemistry/sequencing issues are possible (potential items to modify for next time). That said, for both cases, I would try the items above first since these are where you would need to tune to use this data anyway. Poor quality happens, and people still use the data sometimes, so you are checking that there isn’t bias that could impact the scientific result interpretations! 