Hello, I am currently using the Tuxedo pipeline to identify lncRNAs. I obtained transcript assemblies for each sample using cufflinks, combined them into a single file with cuffcompare, and then extracted lncRNAs based on the class code.

My question is about how to proceed with differential analysis after obtaining the lncRNAs. I intended to create a reference using the combined assembly file produced by cuffcompare and then perform RSEM based on that reference. However, I noticed that when I check the results from RSEM, a single gene is split into multiple parts, and the count values are measured separately.(As you can see in the image below) Do you know why this might be happening?

1. Is it incorrect to create a reference using the cuffcompare result? I'm concerned that if a single gene is split into multiple parts like this, it may pose difficulties in obtaining differentially expressed genes.
2. Are there other good methods for using RSEM after cuffcompare to measure lncRNA expression?
3. Or, are there other effective methods for determining the expression levels of each lncRNA?