The '.' class code shown in the .tracking and .combined.gtf files for multiple input files is a result of the transcripts merged there across multiple samples having different relationships (class code) with the reference transcript. Currently gffcompare uses the uninformative "confused" code . but it would be much more informative if a list of the actual codes encountered there should be used instead, e.g. =c if some of the transcripts on that line have code = and others have code c.
The list should be sorted by frequency -- the most frequent class code should be placed first in that concatenated list.
The '.' class code shown in the .tracking and .combined.gtf files for multiple input files is a result of the transcripts merged there across multiple samples having different relationships (class code) with the reference transcript. Currently gffcompare uses the uninformative "confused" code
.but it would be much more informative if a list of the actual codes encountered there should be used instead, e.g.=cif some of the transcripts on that line have code=and others have codec.The list should be sorted by frequency -- the most frequent class code should be placed first in that concatenated list.