Public Health Alliance for Genomic Epidemiology (PHA4GE) reposted this
Week two of Genomic AMR by Wellcome Connecting Science Learning and Training, we covered databases that contain information on AMR. We learnt about National Center for Biotechnology Information (NCBI) ResFinder and CARD( Comprehensive Antimicrobial Resistance Database). We dived into the way ResFinder was founded and how it works. We also covered, Algorithm tools for detecting AMR : the tools can use either Local alignment search tools, K-mer counting, Hidden Markov Models and Read Mapping among others. One algorithm tool that stood out is Basic Local Alignment Search Tool ( BLAST) that was invented in the 1990s and is still very relevant and useful today. It is especially very important to the NCBI database. There are other relevant tools such as ABRicate, Kleborate for Klebsiella , LRE Finder, pathogen watch and TB-Profiler that is specific for M. tuberculosis and uses read mapping based approach. Many of these databases accept fasta files as the input files. Further, this databases also use other tools that are associated with them to be able to detect AMR such as AMRFinderPlus for NCBI. We also did a comparison and contrast for the three databases NCBI, CARD and ResFinder. It is important to note, that most of these databases are opensource and have been put together through extensive collaborations from all over the world. There is also a hAMRonization tool which was formulated to standadise AMR output tools. This standadisation is being formulated by Public Health Alliance for Genomic Epidemiology (PHA4GE) where the scientists are trying to to develop community standards for more effective use of genomic data in public health. Lastly, we covered data harmonisation which shows that they could be inconsistencies in data due to difference in priorities and expertise. Another challenge is that these tools are limited to AMR related gene presence /absence, not factoring in other AMR variants such as sequence variants, multicopy genes, and/or multi-component systems. This was an interesting week, learning about all these databases and their associated tools and how they work! Such cool concepts no doubt, I cannot wait to contribute some data to these open source tools.