Cuffcompare: difference between "nearest_ref" and "oID"

[wariobrega]

Hello Everyone,

I am a master student new to cufflinks and i am using cuffcompare in order to evaluate an overall analysis on the same samples with and without a GTF Ensembl reference I am performing this for some zebrafish Illumina data I have.

I used cuffcompare with a reference GTF (-r). Although, my combined.gtf have two types of "lines" (meaning isoforms) that differs in the extra field (the one containing all information about gene expresso, exon number, and so on).

The difference is that the extra fields for the former lines looks like: (I will parse the object I am using to store these data)

['gene_id "XLOC_001831"', ' transcript_id "TCONS_00004192"', ' exon_number "12"', ' gene_name "rasgef1bb"', ' oId "ENSDART00000064968"', ' nearest_ref "ENSDART00000064968"', ' class_code "o"', ' tss_id "TSS2969"', '\n']

While the second one looks like:

['gene_id "XLOC_008865"', ' transcript_id "TCONS_00018852"', ' exon_number "2"', ' oId "ENSDART00000014415"', ' class_code "u"', ' tss_id "TSS13648"', '\n']

Apart the missing gene name, I can't understand the difference between "nearest_ref" and oID. Is oID referring to the Cufflinks analysis done using the GTF? and why the oID for the cufflinks run used without GTF (that should look like "CUFF X.X." is not mentoned?

Also, if nearest_ref is referring to the closest GTF annotation given to cuffcompare, why sometimes is different from the oID field since the GTF is the same both with Cufflinks and Cuffcompare? Shouldn't they match?

Thanks in advance for your reply,

Daniele