lapa_correct_talon generates inconsistent corrected abundance and gtf files

[MustafaElshani]

When running the comparison between corrected abundance file and gtf I get the following

 57282 isoforms were quantified.
 50042 isoforms are annotated.
 Only 50042 overlap.
 7240 isoforms quantifed had no corresponding annoation

This is not the case of the parental TALON generated files

Additionally there seems to be some transcript abundance only differing ENST00000268841.10#0 ENST00000268841.10#1
where is this coming from ?

[MustafaElshani]

I think the issue arises due to the way _update_abundace is calculated which takes the linked counts, I think assumption was made that the linked counts would match the way the threshold worked for creation of corrected gtf .

I introduced a further filtering criteria that excludes transcripts in the corrected abundance that are not present in the corrected gtf. I introduced some changes here to match corrected abundance and gtf outputs MustafaElshani@321196e

[MustafaElshani]

Another issue i'm noticing are the improbable start and end coordinates in the output corrected gtf

[MustafaElshani]

Another issue i'm noticing are the improbable start and end coordinates in the output corrected gtf

I think the code logic allows for genes to copies from input gtf without the TES and TSS support and then also creates erroneous exons.