Hello! I am currently using the Cufflinks package to identify lncRNA. I'm curious about the difference in the 'class code' section between the result files of cuffmerge and cuffcompare. As you can see in the picture, cuffmerge's class code consists of only six categories: '=', 'i', 'j', 'u', 'o', and 'x', whereas cuffcompare's class code contains more. What causes this difference? Which file is better to use for identifying long noncoding RNA?