search

gpertea / gffcompare

classify, merge, tracking and annotation of GFF files by comparing to a reference annotation GFF
MIT License
203 stars 32 forks source link

Why the total transcripts in the .tmp file was less than the total transcripts in the input .gtf file #23

Closed syy2017 closed 6 years ago

syy2017 commented 6 years ago

Hi, I am taking a few stringtie assemblies of human single-cell RNAseq data and annotating them back against Gencode v27 gtf files using gffcompare. I have noticed some differences between the total number of transcripts in the one input.gtf and the output file of .tmap after I used the many default modes of gffcompare. The command is used as follows: gffcompare-0.10.1.Linux_x86_64/gffcompare -R -r /ReferenceAnnotation/gencode.v27.annotation.gtf -G
-o gffResultPrefix stringtie_result/SRR.gtf
I have got some information about the differences between the "GffCompare" and "CuffCompare" from the website at https://github.com/gpertea/gffcompare, as follows: Another important difference is that the input transcripts are no longer discarded when they are found to be "intron redundant", i.e. contained within other, longer isoforms. CuffCompare had the -G option to prevent collapsing of such intron redundant isoforms into their longer "containers", but GffCompare has made this the default mode of operation (hence the -G option is no longer needed and is simply ignored when given). Maybe I am missing something obvious, so could you please give advice on how I can run the newest version of gffcompare to preserve the missing transcripts?

gpertea commented 6 years ago

This seems to be related to the long standing bug where GffCompare used to discard "duplicate" transcripts even in "annotation mode". This should be now fixed in release v0.10.2.