faLearn package alongside MIC in your scripts.
As such the package now has less dependencies, and no longer requires compilation.Functions that download data from PATRIC/BV-BRC have now been updated to support changes in BV-BRC ftp.
As of this version the genomic functions of the package have been
deprecated, so that future versions of MIC can focus on MIC analysis.
Any removed functions will be ported to a new package (called faLearn,
see: https://github.com/agerada/faLearn) that focuses on machine learning with
genomic data. The only change for users will be to install and load faLearn.
Functionality added to handle <= and => in MIC values, distinguishing them from
< and >. In general, inhibitory MIC assays report <= and >. The leq and geq
arguments for force_mic can be used to enforce these. The essential_agreement
and compare_mic functions now have tolerate_leq and tolerate_geq arguments.
When all settings are kept at default values, the functions assume typical censoring,
i.e., <= and >.
essential_agreement function to allow a more robust, flexible,
and explicit approach to dealing with censored values. Now, essential_agreement
(and compare_mic) have tolerate_censoring and tolerate_matched_censoring
arguments to control how censored values are handled. The default values should
be appropriate for most situations where the user is comparing an investigational
method to a gold standard method.compare_mic is faster when only one ab is providedsubset S3 method added for mic_validationplot.mic_validation now properly matches dilutions on the lower end of the
scaledroplevels.mic_validation method added that allows unnecessary MIC levels
in a validation object to be droppedtidy_patric_meta_data was missing MICs when laboratory_typing_method was
"MIC" (see 81c69f08)pull_patric_genomes now takes an ab argument to only download strains
where the specified antibiotic was testedAMR v3.0