Hello everybody! I need some conceptual guidance to better understand my results. I am looking for antibiotic resistance (ARGs) genes in shotgun metagenomic reads. I used to approaches:

1) alignment of short reads to ARGs in the CARD database, and 2) upload of open reading frames (predicted from assembled contigs) to GhostKOALA to find antimicrobial resistance signatures based on GHOSTX homology search.

The problem is that the majority of ARGs identified are different depending on the approach. Is that make sense? Are they complementary approaches? Which would be more reliable in terms of functional annotation of metagenome sequences?

Thanks in advance for your time and help!

Pablo