Hello everyone,
I'm using maker since a while now, and I never got any issues about it. I recently tried to include some Augustus prediction inside the annotation process by changing the line : augustus_species=fly as it was specified in some maker tutorial.
At the end of the process, only a dozens of all my contigs (between 500-600), were annotated, and all the rest failed. When I'm looking in the log file, here is what I got :
Gathering GFF3 input into hits - chunk:0
ERROR: Non-unique top level ID for
While this is technically legal in GFF3, it usually
indicates a poorly fomatted GFF3 file (perhaps you
tried to merge two GFF3 files without accounting for
unique IDs). MAKER will not handle these correctly.
--> rank=NA, hostname=GP-Pan
ERROR: Failed while prepare section files
ERROR: Chunk failed at level:12, tier_type:3
FAILED CONTIG:Andromeda
ERROR: Chunk failed at level:4, tier_type:0
FAILED CONTIG:Andromeda
When I tried to look for a solution on maker google groups, the error they got was not exactly the same as me, and the main reasons they suspected the failure was because of too long contig name, which is not my case as it worked without Augustus.
Do you see something obvious that I might have forgot while adding this augustus option ? Thanks for your help !
Roxane
Can you show us a contig name of one that failed ? Then, could you provide as well a piece of the gff3 you use ?
You have an example of name just right above here : Andromeda.
Here some others example of contig names that failed : Draco, Cepheus, Gemini, Contig_5, etc...
Some contigs (a few) succeeded to be annotated. For example Contig_534 etc... Only small contigs.
Are you sure these contig names are unique within the fasta file ?
Yes I'm pretty sure, there is no redundancy. It's the same file I've worked with since the begining. I used it for maker without Augustus. And it was working just fine before. Adding augustus input in the only thing I changed
Ok, so how did you input the augustus data ?
As you adviced me in a previous post, I specified it using augustus_species=fly. I downloaded a fly data set of augustus on thei website, so the set is already trained and ready. It didnt showed off any errors concerning augustus.
If you deactivate/remove it, does it work again properly ?
Hm your guess was right. Deactivating augustus did not changed theses errors. I got exactly the same message. I guess something is wrong with one of my input files.
Remove all the gff3 and add them one by one, doing a run each time (just the time for the tool to parse the gff3 input) to see which one raise the error. Then show us some lines of this file.
It appears that one of my gff was in fact a gtf... Probably I mistake I made while using stringtie output. Problem solved now. Than you for your help !
BTW be careful using stringtie gff3 output, maker doesn't complain but it is not use if it is given as an gff alignment (est_gff or altest_gff).