Potential mapping issues in Cas9 knockout cells with STAR (the truncated RNA)

Question

[Weize]

Hello, In the case of Cas9 knockout (KO) cell, because of the double-strand break in the target exon, it will create something like the "new" splice junctions (shown in the plot below). When building the index and mapping reads using STAR, if still use the original GTF file which represents the wild type, could this lead to potential issues? Specially:

Will truncated RNA fragments fail to map correctly to the modified KO exon (or in other words, the KO gene)? Or even worse, could these fragments be incorrectly mapped to other genes, potentially affecting gene expression quantification? Simple illustration of this KO case:

     -AGTCCTGA-            WT RNA fragment
      ||||||||
-[----AGTCCTGA----]-       WT exon
      ||??????
     -AGGA-               KO RNA fragment, can not align to the WT exon
-[----AGGA-----]-         KO exon (truncated by Cas9 protein)

In the STAR manual, the --outSJfilter* options don’t seem to support my specific case. I’m working with Cas9 knockout (KO) data, and unsure how to properly handle such KO data during alignment. Should I modify the GTF file, adjust specific STAR parameters, or take another approach? Any guidance would be greatly appreciated!

Looking forward to your insights. Thanks!

[rfran010]

STAR will generally be used to align to the genomic reference and the GTF file generally just helps map around splice junctions, but STAR is not restricted to those splice junctions.

"Truncated" RNA fragments should probably not exist as they should be quickly degraded. This could include transcripts with an early stop codon.

In the example you've shown, I think STAR would handle it well as it's designed for splicing. If you had a transcript that copied the "exon" as AGGA, then it's similar to TCCT being spliced out as an intron, since STAR does not rely solely on GTF annotation. However, the sequence is short enough it may just be considered as a deletion. I would need to review how/if STAR handles these differently and if any special parameter would be useful.

All that said, I wouldn't worry about it. The chances of any issue arising are low, but if you really need to analyze this region, then just visually confirm the fragments are mapped correctly on a genome browser. Visualizing the bam or even a signal file with the proper resolution would suffice.

 -NNNNAGTCCTGANNNN-       WT RNA fragment
  ||||||||||||||||        
N[NNNNAGTCCTGANNNN]N      WT exon (genomic)
  ||||||    ||||||
 -NNNNAG----GANNNN-       KO RNA fragment, maps as novel splice junction (read is split)

[Weize]

Thank you so much! I'll check the Cas9 editing site on the genome browser. Initially, I also thought that the "truncated" RNA fragments should be degraded quickly, but I still observed some counts for this truncated gene. But anyway thank you so much for your reply.