Possibly unrelated, but I also had issues using the GFF output from Prokka. My hacky-solution was to edit the attributes field for each sequence ("^contig" in my case) in the GFF file. Example below.

grep "^contig" prokka_out.gff | sed "s/Name=/OtherName=/g" | sed "s/ID=/Name=/g" > amrfinder_out/prokka_out_clean.gff
amrfinder \
--protein prokka_out.faa \
--gff amrfinder_out/prokka_out_clean.gff \
--plus \
--threads 30 > amrfinder_out/amrfinder.tsv

Thanks @julianzaugg!

It appears that a link to instructions our wiki wasn't working. Please let me know if you know where the broken link was coming from. Here's the perl one-liner from the wiki help that does what @julianzaugg's nice sed pipeline does. It also trims off FASTA at the end because we've gotten examples of Prokka GFFs with FASTA files included in them (might not be necessary anymore):

perl -pe '/^##FASTA/ && exit; s/(\W)Name=/$1OldName=/i; s/ID=([^;]+)/ID=$1;Name=$1/' <prokka_output.gff>  > <amrfinder.gff>

hi, I met the same error;
GFF file mismatch.
*** ERROR ***
gff_check.cpp: Protein FASTA id "CRE.p1-1_clust1_00003" is not in the GFF file

But I checked my three input files, all contained the CRE.p1-1_clust1_00003.
gff:
CRE.p1-1_clust1_00003 Prodigal:2.6 CDS 82 486 . + 0 ID=CRE.p1-1_clust1_00003;Name=CRE.p1-1_clust1_00003;Name=tufB;gene=tufB;inference=ab initio prediction:Prodigal:2.6,similar to AA sequence:UniProtKB:P0CE48;locus_tag=CRE.p1-1_clust1_00003;product=Elongation factor Tu 2

faa:

CRE.p1-1_clust1_00003
MFRKLLDEGRAGENVGVLLRGIKREEIERGQVLAKPGSIKPHTQFESEVYILSKDEGGRH
TPFFKGYRPQFYFRTTDVTGTIELPEGVEMVMPGDNVNMVVTLIHPIAMDDGLRFAIREG
GRTVGAGVVAKVIA

ffn:

CRE.p1-1_clust1_00003
ATGTTCCGCAAACTGCTGGACGAAGGCCGTGCTGGTGAGAACGTAGGTGTTCTGCTGCGT
GGTATCAAACGTGAAGAAATCGAACGTGGTCAGGTACTGGCGAAGCCAGGCTCTATCAAG
CCACACACCCAGTTCGAATCTGAAGTGTACATCCTGAGCAAAGATGAAGGTGGTCGTCAC
ACTCCATTCTTCAAAGGCTACCGTCCACAGTTCTACTTCCGTACCACTGACGTGACCGGT
ACCATCGAACTGCCAGAAGGCGTAGAGATGGTAATGCCAGGCGACAACGTGAACATGGTT
GTAACCCTGATTCACCCAATCGCGATGGACGACGGTCTGCGTTTCGCAATCCGTGAAGGC
GGCCGTACCGTTGGCGCAGGTGTTGTTGCTAAAGTTATCGCTTAA

I don't know why, and I hope you can give me some advices，thank you.

We have implemented the amrfinder option

-a ANNOTATION_FORMAT, --annotation_format ANNOTATION_FORMAT
    Type of GFF file: bakta, genbank, microscope, patric, pgap, prokka, pseudomonasdb, rast

What is the format of your GFF file?
Please attach your protein, nucleotide and GFF files for us to test.

Hi @chaoyanggu,

I would also point you at the section of our documentation on input formats. Most likely you just need to use one of the --annotation_format options that Slava mentioned above.

Hope that helps,
Arjun

I added the option -a prokka, now I can running the codes normally, because my gff was generated by prokka, not the default genbank.
Thank you all very much.