Introduction
@@ -147,21 +147,21 @@ make the structure of your data generally look like this:
-2025-09-01
+2025-09-03
abcd
Escherichia coli
S
S
-2025-09-01
+2025-09-03
abcd
Escherichia coli
S
R
-2025-09-01
+2025-09-03
efgh
Escherichia coli
R
diff --git a/articles/datasets.html b/articles/datasets.html
index fce0c3c70..f2acb2e10 100644
--- a/articles/datasets.html
+++ b/articles/datasets.html
@@ -80,7 +80,7 @@
On this page
diff --git a/reference/amr-tidymodels.html b/reference/amr-tidymodels.html
index 1e9cdb02b..b58d86eab 100644
--- a/reference/amr-tidymodels.html
+++ b/reference/amr-tidymodels.html
@@ -179,7 +179,7 @@ may affect the computations for subsequent operations.
#> ✔ broom 1.0.9 ✔ rsample 1.3.1#> ✔ dials 1.4.1 ✔ tibble 3.3.0#> ✔ infer 1.0.9 ✔ tidyr 1.3.1#> ✔ modeldata 1.5.1 ✔ tune 1.3.0#> ✔ modeldata 1.5.1 ✔ tune 2.0.0#> ✔ parsnip 1.3.3 ✔ workflows 1.3.0#> ✔ purrr 1.1.0 ✔ workflowsets 1.1.1#> ✔ recipes 1.3.1 ✔ yardstick 1.3.2#> # A tibble: 4 × 18 #> datetime index method ab_given mo_given host_given input_given#> <dttm> <int> <chr> <chr> <chr> <chr> <chr> #> 1 2025-09-01 15:04:18 1 MIC amoxicillin Escherich… human 8 #> 2 2025-09-01 15:04:19 1 MIC cipro Escherich… human 0.256 #> 3 2025-09-01 15:04:19 1 DISK tobra Escherich… human 16 #> 4 2025-09-01 15:04:19 1 DISK genta Escherich… human 18 #> 1 2025-09-03 09:53:11 1 MIC amoxicillin Escherich… human 8 #> 2 2025-09-03 09:53:11 1 MIC cipro Escherich… human 0.256 #> 3 2025-09-03 09:53:11 1 DISK tobra Escherich… human 16 #> 4 2025-09-03 09:53:12 1 DISK genta Escherich… human 18 #> # ℹ 11 more variables: ab <ab>, mo <mo>, host <chr>, input <chr>, #> # outcome <sir>, notes <chr>, guideline <chr>, ref_table <chr>, uti <lgl>, #> # breakpoint_S_R <chr>, site <chr> #> 1 J3 A 2012-11-21 E. coli R I S S #> 2 R7 A 2018-04-03 K. pneumoniae R I S S #> 3 P3 A 2014-09-19 E. coli R S S S #> 4 P10 A 2015-12-10 E. coli S I S S #> 5 B7 A 2015-03-02 E. coli S S S S #> 6 W3 A 2018-03-31 S. aureus R S R S #> 7 J8 A 2016-06-14 E. coli R S S S #> 8 M3 A 2015-10-25 E. coli R S S S #> 9 J3 A 2019-06-19 E. coli S S S S #> 10 G6 A 2015-04-27 S. aureus S S S S #> # ℹ 2,990 more rows # we will use 'our_data' as the data set name for this tutorial our_data <- example_isolates_unclean"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"taxonomy-of-microorganisms","dir":"Articles","previous_headings":"Preparation","what":"Taxonomy of microorganisms","title":"Conduct AMR data analysis","text":".mo(), users can transform arbitrary microorganism names codes current taxonomy. AMR package contains --date taxonomic data. specific, currently included data retrieved 24 Jun 2024. codes AMR packages come .mo() short, still human readable. importantly, .mo() supports kinds input: first character codes denote taxonomic kingdom, Bacteria (B), Fungi (F), Protozoa (P). AMR package also contain functions directly retrieve taxonomic properties, name, genus, species, family, order, even Gram-stain. start mo_ use .mo() internally, still arbitrary user input can used: Now can thus clean data: Apparently, uncertainty translation taxonomic codes. Let’s check : ’s good.","code":"as.mo(\"Klebsiella pneumoniae\") #> Class 'mo' #> [1] B_KLBSL_PNMN as.mo(\"K. pneumoniae\") #> Class 'mo' #> [1] B_KLBSL_PNMN as.mo(\"KLEPNE\") #> Class 'mo' #> [1] B_KLBSL_PNMN as.mo(\"KLPN\") #> Class 'mo' #> [1] B_KLBSL_PNMN mo_family(\"K. pneumoniae\") #> [1] \"Enterobacteriaceae\" mo_genus(\"K. pneumoniae\") #> [1] \"Klebsiella\" mo_species(\"K. pneumoniae\") #> [1] \"pneumoniae\" mo_gramstain(\"Klebsiella pneumoniae\") #> [1] \"Gram-negative\" mo_ref(\"K. pneumoniae\") #> [1] \"Trevisan, 1887\" mo_snomed(\"K. pneumoniae\") #> [[1]] #> [1] \"1098101000112102\" \"446870005\" \"1098201000112108\" \"409801009\" #> [5] \"56415008\" \"714315002\" \"713926009\" our_data$bacteria <- as.mo(our_data$bacteria, info = TRUE) #> ℹ Retrieved values from the `microorganisms.codes` data set for \"ESCCOL\", #> \"KLEPNE\", \"STAAUR\", and \"STRPNE\". #> ℹ Microorganism translation was uncertain for four microorganisms. Run #> `mo_uncertainties()` to review these uncertainties, or use #> `add_custom_microorganisms()` to add custom entries. mo_uncertainties() #> Matching scores are based on the resemblance between the input and the full #> taxonomic name, and the pathogenicity in humans. See `?mo_matching_score`. #> Colour keys: 0.000-0.549 0.550-0.649 0.650-0.749 0.750-1.000 #> #> -------------------------------------------------------------------------------- #> \"E. coli\" -> Escherichia coli (B_ESCHR_COLI, 0.688) #> Also matched: Enterococcus crotali (0.650), Escherichia coli coli #> (0.643), Escherichia coli expressing (0.611), Enterobacter cowanii #> (0.600), Enterococcus columbae (0.595), Enterococcus camelliae (0.591), #> Enterococcus casseliflavus (0.577), Enterobacter cloacae cloacae #> (0.571), Enterobacter cloacae complex (0.571), and Enterobacter cloacae #> dissolvens (0.565) #> -------------------------------------------------------------------------------- #> \"K. pneumoniae\" -> Klebsiella pneumoniae (B_KLBSL_PNMN, 0.786) #> Also matched: Klebsiella pneumoniae complex (0.707), Klebsiella #> pneumoniae ozaenae (0.707), Klebsiella pneumoniae pneumoniae (0.688), #> Klebsiella pneumoniae rhinoscleromatis (0.658), Klebsiella pasteurii #> (0.500), Klebsiella planticola (0.500), Kingella potus (0.400), #> Kluyveromyces pseudotropicale (0.386), Kluyveromyces pseudotropicalis #> (0.363), and Kosakonia pseudosacchari (0.361) #> -------------------------------------------------------------------------------- #> \"S. aureus\" -> Staphylococcus aureus (B_STPHY_AURS, 0.690) #> Also matched: Staphylococcus aureus aureus (0.643), Staphylococcus #> argenteus (0.625), Staphylococcus aureus anaerobius (0.625), #> Staphylococcus auricularis (0.615), Salmonella Aurelianis (0.595), #> Salmonella Aarhus (0.588), Salmonella Amounderness (0.587), #> Staphylococcus argensis (0.587), Streptococcus australis (0.587), and #> Salmonella choleraesuis arizonae (0.562) #> -------------------------------------------------------------------------------- #> \"S. pneumoniae\" -> Streptococcus pneumoniae (B_STRPT_PNMN, 0.750) #> Also matched: Streptococcus pseudopneumoniae (0.700), Streptococcus #> phocae salmonis (0.552), Serratia proteamaculans quinovora (0.545), #> Streptococcus pseudoporcinus (0.536), Staphylococcus piscifermentans #> (0.533), Staphylococcus pseudintermedius (0.532), Serratia #> proteamaculans proteamaculans (0.526), Streptococcus gallolyticus #> pasteurianus (0.526), Salmonella Portanigra (0.524), and Streptococcus #> periodonticum (0.519) #> #> Only the first 10 other matches of each record are shown. Run #> `print(mo_uncertainties(), n = ...)` to view more entries, or save #> `mo_uncertainties()` to an object."},{"path":"https://amr-for-r.org/articles/AMR.html","id":"antibiotic-results","dir":"Articles","previous_headings":"Preparation","what":"Antibiotic results","title":"Conduct AMR data analysis","text":"column antibiotic test results must also cleaned. AMR package comes three new data types work test results: mic minimal inhibitory concentrations (MIC), disk disk diffusion diameters, sir SIR data interpreted already. package can also determine SIR values based MIC disk diffusion values, read .sir() page. now, just clean SIR columns data using dplyr: basically cleaning, time start data inclusion.","code":"# method 1, be explicit about the columns: our_data <- our_data %>% mutate_at(vars(AMX:GEN), as.sir) # method 2, let the AMR package determine the eligible columns our_data <- our_data %>% mutate_if(is_sir_eligible, as.sir) # result: our_data #> # A tibble: 3,000 × 8 #> patient_id hospital date bacteria AMX AMC CIP GEN #> #> 1 J3 A 2012-11-21 B_ESCHR_COLI R I S S #> 2 R7 A 2018-04-03 B_KLBSL_PNMN R I S S #> 3 P3 A 2014-09-19 B_ESCHR_COLI R S S S #> 4 P10 A 2015-12-10 B_ESCHR_COLI S I S S #> 5 B7 A 2015-03-02 B_ESCHR_COLI S S S S #> 6 W3 A 2018-03-31 B_STPHY_AURS R S R S #> 7 J8 A 2016-06-14 B_ESCHR_COLI R S S S #> 8 M3 A 2015-10-25 B_ESCHR_COLI R S S S #> 9 J3 A 2019-06-19 B_ESCHR_COLI S S S S #> 10 G6 A 2015-04-27 B_STPHY_AURS S S S S #> # ℹ 2,990 more rows"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"first-isolates","dir":"Articles","previous_headings":"Preparation","what":"First isolates","title":"Conduct AMR data analysis","text":"need know isolates can actually use analysis without repetition bias. conduct analysis antimicrobial resistance, must include first isolate every patient per episode (Hindler et al., Clin Infect Dis. 2007). , easily get overestimate underestimate resistance antibiotic. Imagine patient admitted MRSA found 5 different blood cultures following weeks (yes, countries like Netherlands blood drawing policies). resistance percentage oxacillin isolates overestimated, included MRSA . clearly selection bias. Clinical Laboratory Standards Institute (CLSI) appoints follows: (…) preparing cumulative antibiogram guide clinical decisions empirical antimicrobial therapy initial infections, first isolate given species per patient, per analysis period (eg, one year) included, irrespective body site, antimicrobial susceptibility profile, phenotypical characteristics (eg, biotype). first isolate easily identified, cumulative antimicrobial susceptibility test data prepared using first isolate generally comparable cumulative antimicrobial susceptibility test data calculated methods, providing duplicate isolates excluded. M39-A4 Analysis Presentation Cumulative Antimicrobial Susceptibility Test Data, 4th Edition. CLSI, 2014. Chapter 6.4 AMR package includes methodology first_isolate() function able apply four different methods defined Hindler et al. 2007: phenotype-based, episode-based, patient-based, isolate-based. right method depends goals analysis, default phenotype-based method case method properly correct duplicate isolates. Read methods first_isolate() page. outcome function can easily added data: 91% suitable resistance analysis! can now filter filter() function, also dplyr package: future use, two syntaxes can shortened: end 2 724 isolates analysis. Now data looks like: Time analysis.","code":"our_data <- our_data %>% mutate(first = first_isolate(info = TRUE)) #> ℹ Determining first isolates using an episode length of 365 days #> ℹ Using column 'bacteria' as input for `col_mo`. #> ℹ Using column 'date' as input for `col_date`. #> ℹ Using column 'patient_id' as input for `col_patient_id`. #> ℹ Basing inclusion on all antimicrobial results, using a points threshold #> of 2 #> => Found 2,724 'phenotype-based' first isolates (90.8% of total where a #> microbial ID was available) our_data_1st <- our_data %>% filter(first == TRUE) our_data_1st <- our_data %>% filter_first_isolate() our_data_1st #> # A tibble: 2,724 × 9 #> patient_id hospital date bacteria AMX AMC CIP GEN first #> #> 1 J3 A 2012-11-21 B_ESCHR_COLI R I S S TRUE #> 2 R7 A 2018-04-03 B_KLBSL_PNMN R I S S TRUE #> 3 P3 A 2014-09-19 B_ESCHR_COLI R S S S TRUE #> 4 P10 A 2015-12-10 B_ESCHR_COLI S I S S TRUE #> 5 B7 A 2015-03-02 B_ESCHR_COLI S S S S TRUE #> 6 W3 A 2018-03-31 B_STPHY_AURS R S R S TRUE #> 7 M3 A 2015-10-25 B_ESCHR_COLI R S S S TRUE #> 8 J3 A 2019-06-19 B_ESCHR_COLI S S S S TRUE #> 9 G6 A 2015-04-27 B_STPHY_AURS S S S S TRUE #> 10 P4 A 2011-06-21 B_ESCHR_COLI S S S S TRUE #> # ℹ 2,714 more rows"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"analysing-the-data","dir":"Articles","previous_headings":"","what":"Analysing the data","title":"Conduct AMR data analysis","text":"base R summary() function gives good first impression, comes support new mo sir classes now data set:","code":"summary(our_data_1st) #> patient_id hospital date #> Length:2724 Length:2724 Min. :2011-01-01 #> Class :character Class :character 1st Qu.:2013-04-07 #> Mode :character Mode :character Median :2015-06-03 #> Mean :2015-06-09 #> 3rd Qu.:2017-08-11 #> Max. :2019-12-27 #> bacteria AMX AMC #> Class :mo Class:sir Class:sir #> :0 %S :41.6% (n=1133) %S :52.6% (n=1432) #> Unique:4 %SDD : 0.0% (n=0) %SDD : 0.0% (n=0) #> #1 :B_ESCHR_COLI %I :16.4% (n=446) %I :12.2% (n=333) #> #2 :B_STPHY_AURS %R :42.0% (n=1145) %R :35.2% (n=959) #> #3 :B_STRPT_PNMN %NI : 0.0% (n=0) %NI : 0.0% (n=0) #> CIP GEN first #> Class:sir Class:sir Mode:logical #> %S :52.5% (n=1431) %S :61.0% (n=1661) TRUE:2724 #> %SDD : 0.0% (n=0) %SDD : 0.0% (n=0) #> %I : 6.5% (n=176) %I : 3.0% (n=82) #> %R :41.0% (n=1117) %R :36.0% (n=981) #> %NI : 0.0% (n=0) %NI : 0.0% (n=0) glimpse(our_data_1st) #> Rows: 2,724 #> Columns: 9 #> $ patient_id \"J3\", \"R7\", \"P3\", \"P10\", \"B7\", \"W3\", \"M3\", \"J3\", \"G6\", \"P4\"… #> $ hospital \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\", \"A\",… #> $ date 2012-11-21, 2018-04-03, 2014-09-19, 2015-12-10, 2015-03-02… #> $ bacteria \"B_ESCHR_COLI\", \"B_KLBSL_PNMN\", \"B_ESCHR_COLI\", \"B_ESCHR_COL… #> $ AMX R, R, R, S, S, R, R, S, S, S, S, R, S, S, R, R, R, R, S, R,… #> $ AMC I, I, S, I, S, S, S, S, S, S, S, S, S, S, S, S, S, R, S, S,… #> $ CIP S, S, S, S, S, R, S, S, S, S, S, S, S, S, S, S, S, S, S, S,… #> $ GEN S, S, S, S, S, S, S, S, S, S, S, R, S, S, S, S, S, S, S, S,… #> $ first TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,… # number of unique values per column: sapply(our_data_1st, n_distinct) #> patient_id hospital date bacteria AMX AMC CIP #> 260 3 1854 4 3 3 3 #> GEN first #> 3 1"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"availability-of-species","dir":"Articles","previous_headings":"Analysing the data","what":"Availability of species","title":"Conduct AMR data analysis","text":"just get idea species distributed, create frequency table count() based name microorganisms:","code":"our_data %>% count(mo_name(bacteria), sort = TRUE) #> # A tibble: 4 × 2 #> `mo_name(bacteria)` n #> #> 1 Escherichia coli 1518 #> 2 Staphylococcus aureus 730 #> 3 Streptococcus pneumoniae 426 #> 4 Klebsiella pneumoniae 326 our_data_1st %>% count(mo_name(bacteria), sort = TRUE) #> # A tibble: 4 × 2 #> `mo_name(bacteria)` n #> #> 1 Escherichia coli 1321 #> 2 Staphylococcus aureus 682 #> 3 Streptococcus pneumoniae 402 #> 4 Klebsiella pneumoniae 319"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"select-and-filter-with-antibiotic-selectors","dir":"Articles","previous_headings":"Analysing the data","what":"Select and filter with antibiotic selectors","title":"Conduct AMR data analysis","text":"Using -called antibiotic class selectors, can select filter columns based antibiotic class antibiotic results :","code":"our_data_1st %>% select(date, aminoglycosides()) #> ℹ For `aminoglycosides()` using column 'GEN' (gentamicin) #> # A tibble: 2,724 × 2 #> date GEN #> #> 1 2012-11-21 S #> 2 2018-04-03 S #> 3 2014-09-19 S #> 4 2015-12-10 S #> 5 2015-03-02 S #> 6 2018-03-31 S #> 7 2015-10-25 S #> 8 2019-06-19 S #> 9 2015-04-27 S #> 10 2011-06-21 S #> # ℹ 2,714 more rows our_data_1st %>% select(bacteria, betalactams()) #> ℹ For `betalactams()` using columns 'AMX' (amoxicillin) and 'AMC' #> (amoxicillin/clavulanic acid) #> # A tibble: 2,724 × 3 #> bacteria AMX AMC #> #> 1 B_ESCHR_COLI R I #> 2 B_KLBSL_PNMN R I #> 3 B_ESCHR_COLI R S #> 4 B_ESCHR_COLI S I #> 5 B_ESCHR_COLI S S #> 6 B_STPHY_AURS R S #> 7 B_ESCHR_COLI R S #> 8 B_ESCHR_COLI S S #> 9 B_STPHY_AURS S S #> 10 B_ESCHR_COLI S S #> # ℹ 2,714 more rows our_data_1st %>% select(bacteria, where(is.sir)) #> # A tibble: 2,724 × 5 #> bacteria AMX AMC CIP GEN #> #> 1 B_ESCHR_COLI R I S S #> 2 B_KLBSL_PNMN R I S S #> 3 B_ESCHR_COLI R S S S #> 4 B_ESCHR_COLI S I S S #> 5 B_ESCHR_COLI S S S S #> 6 B_STPHY_AURS R S R S #> 7 B_ESCHR_COLI R S S S #> 8 B_ESCHR_COLI S S S S #> 9 B_STPHY_AURS S S S S #> 10 B_ESCHR_COLI S S S S #> # ℹ 2,714 more rows # filtering using AB selectors is also possible: our_data_1st %>% filter(any(aminoglycosides() == \"R\")) #> ℹ For `aminoglycosides()` using column 'GEN' (gentamicin) #> # A tibble: 981 × 9 #> patient_id hospital date bacteria AMX AMC CIP GEN first #> #> 1 J5 A 2017-12-25 B_STRPT_PNMN R S S R TRUE #> 2 X1 A 2017-07-04 B_STPHY_AURS R S S R TRUE #> 3 B3 A 2016-07-24 B_ESCHR_COLI S S S R TRUE #> 4 V7 A 2012-04-03 B_ESCHR_COLI S S S R TRUE #> 5 C9 A 2017-03-23 B_ESCHR_COLI S S S R TRUE #> 6 R1 A 2018-06-10 B_STPHY_AURS S S S R TRUE #> 7 S2 A 2013-07-19 B_STRPT_PNMN S S S R TRUE #> 8 P5 A 2019-03-09 B_STPHY_AURS S S S R TRUE #> 9 Q8 A 2019-08-10 B_STPHY_AURS S S S R TRUE #> 10 K5 A 2013-03-15 B_STRPT_PNMN S S S R TRUE #> # ℹ 971 more rows our_data_1st %>% filter(all(betalactams() == \"R\")) #> ℹ For `betalactams()` using columns 'AMX' (amoxicillin) and 'AMC' #> (amoxicillin/clavulanic acid) #> # A tibble: 462 × 9 #> patient_id hospital date bacteria AMX AMC CIP GEN first #> #> 1 M7 A 2013-07-22 B_STRPT_PNMN R R S S TRUE #> 2 R10 A 2013-12-20 B_STPHY_AURS R R S S TRUE #> 3 R7 A 2015-10-25 B_STPHY_AURS R R S S TRUE #> 4 R8 A 2019-10-25 B_STPHY_AURS R R S S TRUE #> 5 B6 A 2016-11-20 B_ESCHR_COLI R R R R TRUE #> 6 I7 A 2015-08-19 B_ESCHR_COLI R R S S TRUE #> 7 N3 A 2014-12-29 B_STRPT_PNMN R R R S TRUE #> 8 Q2 A 2019-09-22 B_ESCHR_COLI R R S S TRUE #> 9 X7 A 2011-03-20 B_ESCHR_COLI R R S R TRUE #> 10 V1 A 2018-08-07 B_STPHY_AURS R R S S TRUE #> # ℹ 452 more rows # even works in base R (since R 3.0): our_data_1st[all(betalactams() == \"R\"), ] #> ℹ For `betalactams()` using columns 'AMX' (amoxicillin) and 'AMC' #> (amoxicillin/clavulanic acid) #> # A tibble: 462 × 9 #> patient_id hospital date bacteria AMX AMC CIP GEN first #> #> 1 M7 A 2013-07-22 B_STRPT_PNMN R R S S TRUE #> 2 R10 A 2013-12-20 B_STPHY_AURS R R S S TRUE #> 3 R7 A 2015-10-25 B_STPHY_AURS R R S S TRUE #> 4 R8 A 2019-10-25 B_STPHY_AURS R R S S TRUE #> 5 B6 A 2016-11-20 B_ESCHR_COLI R R R R TRUE #> 6 I7 A 2015-08-19 B_ESCHR_COLI R R S S TRUE #> 7 N3 A 2014-12-29 B_STRPT_PNMN R R R S TRUE #> 8 Q2 A 2019-09-22 B_ESCHR_COLI R R S S TRUE #> 9 X7 A 2011-03-20 B_ESCHR_COLI R R S R TRUE #> 10 V1 A 2018-08-07 B_STPHY_AURS R R S S TRUE #> # ℹ 452 more rows"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"generate-antibiograms","dir":"Articles","previous_headings":"Analysing the data","what":"Generate antibiograms","title":"Conduct AMR data analysis","text":"Since AMR v2.0 (March 2023), easy create different types antibiograms, support 20 different languages. four antibiogram types, proposed Klinker et al. (2021, DOI 10.1177/20499361211011373), supported new antibiogram() function: Traditional Antibiogram (TA) e.g, susceptibility Pseudomonas aeruginosa piperacillin/tazobactam (TZP) Combination Antibiogram (CA) e.g, sdditional susceptibility Pseudomonas aeruginosa TZP + tobramycin versus TZP alone Syndromic Antibiogram (SA) e.g, susceptibility Pseudomonas aeruginosa TZP among respiratory specimens (obtained among ICU patients ) Weighted-Incidence Syndromic Combination Antibiogram (WISCA) e.g, susceptibility Pseudomonas aeruginosa TZP among respiratory specimens (obtained among ICU patients ) male patients age >=65 years heart failure section, show use antibiogram() function create antibiogram types. starters, included example_isolates data set looks like:","code":"example_isolates #> # A tibble: 2,000 × 46 #> date patient age gender ward mo PEN OXA FLC AMX #> #> 1 2002-01-02 A77334 65 F Clinical B_ESCHR_COLI R NA NA NA #> 2 2002-01-03 A77334 65 F Clinical B_ESCHR_COLI R NA NA NA #> 3 2002-01-07 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 4 2002-01-07 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 5 2002-01-13 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 6 2002-01-13 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 7 2002-01-14 462729 78 M Clinical B_STPHY_AURS R NA S R #> 8 2002-01-14 462729 78 M Clinical B_STPHY_AURS R NA S R #> 9 2002-01-16 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 10 2002-01-17 858515 79 F ICU B_STPHY_EPDR R NA S NA #> # ℹ 1,990 more rows #> # ℹ 36 more variables: AMC , AMP , TZP , CZO , FEP , #> # CXM , FOX , CTX , CAZ , CRO , GEN , #> # TOB , AMK , KAN , TMP , SXT , NIT , #> # FOS , LNZ , CIP , MFX , VAN , TEC , #> # TCY , TGC , DOX , ERY , CLI , AZM , #> # IPM , MEM , MTR , CHL , COL , MUP , …"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"traditional-antibiogram","dir":"Articles","previous_headings":"Analysing the data > Generate antibiograms","what":"Traditional Antibiogram","title":"Conduct AMR data analysis","text":"create traditional antibiogram, simply state antibiotics used. antibiotics argument antibiogram() function supports (combination) previously mentioned antibiotic class selectors: Notice antibiogram() function automatically prints right format using Quarto R Markdown (page), even applies italics taxonomic names (using italicise_taxonomy() internally). also uses language OS either English, Arabic, Bengali, Chinese, Czech, Danish, Dutch, Finnish, French, German, Greek, Hindi, Indonesian, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Romanian, Russian, Spanish, Swahili, Swedish, Turkish, Ukrainian, Urdu, Vietnamese. next example, force language Spanish using language argument:","code":"antibiogram(example_isolates, antibiotics = c(aminoglycosides(), carbapenems())) #> ℹ For `aminoglycosides()` using columns 'GEN' (gentamicin), 'TOB' #> (tobramycin), 'AMK' (amikacin), and 'KAN' (kanamycin) #> ℹ For `carbapenems()` using columns 'IPM' (imipenem) and 'MEM' (meropenem) antibiogram(example_isolates, mo_transform = \"gramstain\", antibiotics = aminoglycosides(), ab_transform = \"name\", language = \"es\") #> ℹ For `aminoglycosides()` using columns 'GEN' (gentamicin), 'TOB' #> (tobramycin), 'AMK' (amikacin), and 'KAN' (kanamycin)"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"combined-antibiogram","dir":"Articles","previous_headings":"Analysing the data > Generate antibiograms","what":"Combined Antibiogram","title":"Conduct AMR data analysis","text":"create combined antibiogram, use antibiotic codes names plus + character like :","code":"combined_ab <- antibiogram(example_isolates, antibiotics = c(\"TZP\", \"TZP+TOB\", \"TZP+GEN\"), ab_transform = NULL) combined_ab"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"syndromic-antibiogram","dir":"Articles","previous_headings":"Analysing the data > Generate antibiograms","what":"Syndromic Antibiogram","title":"Conduct AMR data analysis","text":"create syndromic antibiogram, syndromic_group argument must used. can column data, e.g. ifelse() calculations based certain columns:","code":"antibiogram(example_isolates, antibiotics = c(aminoglycosides(), carbapenems()), syndromic_group = \"ward\") #> ℹ For `aminoglycosides()` using columns 'GEN' (gentamicin), 'TOB' #> (tobramycin), 'AMK' (amikacin), and 'KAN' (kanamycin) #> ℹ For `carbapenems()` using columns 'IPM' (imipenem) and 'MEM' (meropenem)"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"weighted-incidence-syndromic-combination-antibiogram-wisca","dir":"Articles","previous_headings":"Analysing the data > Generate antibiograms","what":"Weighted-Incidence Syndromic Combination Antibiogram (WISCA)","title":"Conduct AMR data analysis","text":"create Weighted-Incidence Syndromic Combination Antibiogram (WISCA), simply set wisca = TRUE antibiogram() function, use dedicated wisca() function. Unlike traditional antibiograms, WISCA provides syndrome-based susceptibility estimates, weighted pathogen incidence antimicrobial susceptibility patterns. WISCA uses Bayesian decision model integrate data multiple pathogens, improving empirical therapy guidance, especially low-incidence infections. pathogen-agnostic, meaning results syndrome-based rather stratified microorganism. reliable results, ensure data includes first isolates (use first_isolate()) consider filtering top n species (use top_n_microorganisms()), WISCA outcomes meaningful based robust incidence estimates. patient- syndrome-specific WISCA, run function grouped tibble, .e., using group_by() first:","code":"example_isolates %>% wisca(antibiotics = c(\"TZP\", \"TZP+TOB\", \"TZP+GEN\"), minimum = 10) # Recommended threshold: ≥30 example_isolates %>% top_n_microorganisms(n = 10) %>% group_by(age_group = age_groups(age, c(25, 50, 75)), gender) %>% wisca(antibiotics = c(\"TZP\", \"TZP+TOB\", \"TZP+GEN\"))"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"plotting-antibiograms","dir":"Articles","previous_headings":"Analysing the data > Generate antibiograms","what":"Plotting antibiograms","title":"Conduct AMR data analysis","text":"Antibiograms can plotted using autoplot() ggplot2 packages, since AMR package provides extension function: calculate antimicrobial resistance sensible way, also correcting results, use resistance() susceptibility() functions.","code":"autoplot(combined_ab)"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"resistance-percentages","dir":"Articles","previous_headings":"Analysing the data","what":"Resistance percentages","title":"Conduct AMR data analysis","text":"functions resistance() susceptibility() can used calculate antimicrobial resistance susceptibility. specific analyses, functions proportion_S(), proportion_SI(), proportion_I(), proportion_IR() proportion_R() can used determine proportion specific antimicrobial outcome. functions contain minimum argument, denoting minimum required number test results returning value. functions otherwise return NA. default minimum = 30, following CLSI M39-A4 guideline applying microbial epidemiology. per EUCAST guideline 2019, calculate resistance proportion R (proportion_R(), equal resistance()) susceptibility proportion S (proportion_SI(), equal susceptibility()). functions can used : can used conjunction group_by() summarise(), dplyr package:","code":"our_data_1st %>% resistance(AMX) #> [1] 0.4203377 our_data_1st %>% group_by(hospital) %>% summarise(amoxicillin = resistance(AMX)) #> # A tibble: 3 × 2 #> hospital amoxicillin #> #> 1 A 0.340 #> 2 B 0.551 #> 3 C 0.370"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"interpreting-mic-and-disk-diffusion-values","dir":"Articles","previous_headings":"Analysing the data","what":"Interpreting MIC and Disk Diffusion Values","title":"Conduct AMR data analysis","text":"Minimal inhibitory concentration (MIC) values disk diffusion diameters can interpreted clinical breakpoints (SIR) using .sir(). ’s example randomly generated MIC values Klebsiella pneumoniae ciprofloxacin: allows direct interpretation according EUCAST CLSI breakpoints, facilitating automated AMR data processing.","code":"set.seed(123) mic_values <- random_mic(100) sir_values <- as.sir(mic_values, mo = \"K. pneumoniae\", ab = \"cipro\", guideline = \"EUCAST 2024\") my_data <- tibble(MIC = mic_values, SIR = sir_values) my_data #> # A tibble: 100 × 2 #> MIC SIR #> #> 1 <=0.0001 S #> 2 0.0160 S #> 3 >=8.0000 R #> 4 0.0320 S #> 5 0.0080 S #> 6 64.0000 R #> 7 0.0080 S #> 8 0.1250 S #> 9 0.0320 S #> 10 0.0002 S #> # ℹ 90 more rows"},{"path":"https://amr-for-r.org/articles/AMR.html","id":"plotting-mic-and-sir-interpretations","dir":"Articles","previous_headings":"Analysing the data","what":"Plotting MIC and SIR Interpretations","title":"Conduct AMR data analysis","text":"can visualise MIC distributions SIR interpretations using ggplot2, using new scale_y_mic() y-axis scale_colour_sir() colour-code SIR categories. plot provides intuitive way assess susceptibility patterns across different groups incorporating clinical breakpoints. straightforward less manual approach, ggplot2’s function autoplot() extended package directly plot MIC disk diffusion values: Author: Dr. Matthijs Berends, 23rd Feb 2025","code":"# add a group my_data$group <- rep(c(\"A\", \"B\", \"C\", \"D\"), each = 25) ggplot(my_data, aes(x = group, y = MIC, colour = SIR)) + geom_jitter(width = 0.2, size = 2) + geom_boxplot(fill = NA, colour = \"grey40\") + scale_y_mic() + scale_colour_sir() + labs(title = \"MIC Distribution and SIR Interpretation\", x = \"Sample Groups\", y = \"MIC (mg/L)\") autoplot(mic_values) # by providing `mo` and `ab`, colours will indicate the SIR interpretation: autoplot(mic_values, mo = \"K. pneumoniae\", ab = \"cipro\", guideline = \"EUCAST 2024\")"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"introduction","dir":"Articles","previous_headings":"","what":"Introduction","title":"AMR for Python","text":"AMR package R powerful tool antimicrobial resistance (AMR) analysis. provides extensive features handling microbial antimicrobial data. However, work primarily Python, now intuitive option available: AMR Python package. Python package wrapper around AMR R package. uses rpy2 package internally. Despite need R installed, Python users can now easily work AMR data directly Python code.","code":""},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"prerequisites","dir":"Articles","previous_headings":"","what":"Prerequisites","title":"AMR for Python","text":"package tested virtual environment (venv). can set environment running: can activate environment, venv ready work .","code":"# linux and macOS: python -m venv /path/to/new/virtual/environment # Windows: python -m venv C:\\path\\to\\new\\virtual\\environment"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"install-amr","dir":"Articles","previous_headings":"","what":"Install AMR","title":"AMR for Python","text":"Since Python package available official Python Package Index, can just run: Make sure R installed. need install AMR R package, installed automatically. Linux: macOS (using Homebrew): Windows, visit CRAN download page download install R.","code":"pip install AMR # Ubuntu / Debian sudo apt install r-base # Fedora: sudo dnf install R # CentOS/RHEL sudo yum install R brew install r"},{"path":[]},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"cleaning-taxonomy","dir":"Articles","previous_headings":"Examples of Usage","what":"Cleaning Taxonomy","title":"AMR for Python","text":"’s example demonstrates clean microorganism drug names using AMR Python package:","code":"import pandas as pd import AMR # Sample data data = { \"MOs\": ['E. coli', 'ESCCOL', 'esco', 'Esche coli'], \"Drug\": ['Cipro', 'CIP', 'J01MA02', 'Ciproxin'] } df = pd.DataFrame(data) # Use AMR functions to clean microorganism and drug names df['MO_clean'] = AMR.mo_name(df['MOs']) df['Drug_clean'] = AMR.ab_name(df['Drug']) # Display the results print(df)"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"explanation","dir":"Articles","previous_headings":"Examples of Usage > Cleaning Taxonomy","what":"Explanation","title":"AMR for Python","text":"mo_name: function standardises microorganism names. , different variations Escherichia coli (“E. coli”, “ESCCOL”, “esco”, “Esche coli”) converted correct, standardised form, “Escherichia coli”. ab_name: Similarly, function standardises antimicrobial names. different representations ciprofloxacin (e.g., “Cipro”, “CIP”, “J01MA02”, “Ciproxin”) converted standard name, “Ciprofloxacin”.","code":""},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"calculating-amr","dir":"Articles","previous_headings":"Examples of Usage","what":"Calculating AMR","title":"AMR for Python","text":"","code":"import AMR import pandas as pd df = AMR.example_isolates result = AMR.resistance(df[\"AMX\"]) print(result) [0.59555556]"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"generating-antibiograms","dir":"Articles","previous_headings":"Examples of Usage","what":"Generating Antibiograms","title":"AMR for Python","text":"One core functions AMR package generating antibiogram, table summarises antimicrobial susceptibility bacterial isolates. ’s can generate antibiogram Python: example, generate antibiogram selecting various antibiotics.","code":"result2a = AMR.antibiogram(df[[\"mo\", \"AMX\", \"CIP\", \"TZP\"]]) print(result2a) result2b = AMR.antibiogram(df[[\"mo\", \"AMX\", \"CIP\", \"TZP\"]], mo_transform = \"gramstain\") print(result2b)"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"taxonomic-data-sets-now-in-python","dir":"Articles","previous_headings":"Examples of Usage","what":"Taxonomic Data Sets Now in Python!","title":"AMR for Python","text":"Python user, might like important data sets AMR R package, microorganisms, antimicrobials, clinical_breakpoints, example_isolates, now available regular Python data frames:","code":"AMR.microorganisms AMR.antimicrobials"},{"path":"https://amr-for-r.org/articles/AMR_for_Python.html","id":"conclusion","dir":"Articles","previous_headings":"","what":"Conclusion","title":"AMR for Python","text":"AMR Python package, Python users can now effortlessly call R functions AMR R package. eliminates need complex rpy2 configurations provides clean, easy--use interface antimicrobial resistance analysis. examples provided demonstrate can applied typical workflows, standardising microorganism antimicrobial names calculating resistance. just running import AMR, users can seamlessly integrate robust features R AMR package Python workflows. Whether ’re cleaning data analysing resistance patterns, AMR Python package makes easy work AMR data Python.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"example-1-using-antimicrobial-selectors","dir":"Articles","previous_headings":"","what":"Example 1: Using Antimicrobial Selectors","title":"AMR with tidymodels","text":"leveraging power tidymodels AMR package, ’ll build reproducible machine learning workflow predict Gramstain microorganism two important antibiotic classes: aminoglycosides beta-lactams.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"objective","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors","what":"Objective","title":"AMR with tidymodels","text":"goal build predictive model using tidymodels framework determine Gramstain microorganism based microbial data. : Preprocess data using selector functions aminoglycosides() betalactams(). Define logistic regression model prediction. Use structured tidymodels workflow preprocess, train, evaluate model.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"data-preparation","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors","what":"Data Preparation","title":"AMR with tidymodels","text":"begin loading required libraries preparing example_isolates dataset AMR package. Prepare data: Explanation: aminoglycosides() betalactams() dynamically select columns antimicrobials classes. drop_na() ensures model receives complete cases training.","code":"# Load required libraries library(AMR) # For AMR data analysis library(tidymodels) # For machine learning workflows, and data manipulation (dplyr, tidyr, ...) # Your data could look like this: example_isolates #> # A tibble: 2,000 × 46 #> date patient age gender ward mo PEN OXA FLC AMX #> #> 1 2002-01-02 A77334 65 F Clinical B_ESCHR_COLI R NA NA NA #> 2 2002-01-03 A77334 65 F Clinical B_ESCHR_COLI R NA NA NA #> 3 2002-01-07 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 4 2002-01-07 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 5 2002-01-13 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 6 2002-01-13 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 7 2002-01-14 462729 78 M Clinical B_STPHY_AURS R NA S R #> 8 2002-01-14 462729 78 M Clinical B_STPHY_AURS R NA S R #> 9 2002-01-16 067927 45 F ICU B_STPHY_EPDR R NA R NA #> 10 2002-01-17 858515 79 F ICU B_STPHY_EPDR R NA S NA #> # ℹ 1,990 more rows #> # ℹ 36 more variables: AMC , AMP , TZP , CZO , FEP , #> # CXM , FOX , CTX , CAZ , CRO , GEN , #> # TOB , AMK , KAN , TMP , SXT , NIT , #> # FOS , LNZ , CIP , MFX , VAN , TEC , #> # TCY , TGC , DOX , ERY , CLI , AZM , #> # IPM , MEM , MTR , CHL , COL , MUP , … # Select relevant columns for prediction data <- example_isolates %>% # select AB results dynamically select(mo, aminoglycosides(), betalactams()) %>% # replace NAs with NI (not-interpretable) mutate(across(where(is.sir), ~replace_na(.x, \"NI\")), # make factors of SIR columns across(where(is.sir), as.integer), # get Gramstain of microorganisms mo = as.factor(mo_gramstain(mo))) %>% # drop NAs - the ones without a Gramstain (fungi, etc.) drop_na() #> ℹ For `aminoglycosides()` using columns 'GEN' (gentamicin), 'TOB' #> (tobramycin), 'AMK' (amikacin), and 'KAN' (kanamycin) #> ℹ For `betalactams()` using columns 'PEN' (benzylpenicillin), 'OXA' #> (oxacillin), 'FLC' (flucloxacillin), 'AMX' (amoxicillin), 'AMC' #> (amoxicillin/clavulanic acid), 'AMP' (ampicillin), 'TZP' #> (piperacillin/tazobactam), 'CZO' (cefazolin), 'FEP' (cefepime), 'CXM' #> (cefuroxime), 'FOX' (cefoxitin), 'CTX' (cefotaxime), 'CAZ' (ceftazidime), #> 'CRO' (ceftriaxone), 'IPM' (imipenem), and 'MEM' (meropenem)"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"defining-the-workflow","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors","what":"Defining the Workflow","title":"AMR with tidymodels","text":"now define tidymodels workflow, consists three steps: preprocessing, model specification, fitting.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"preprocessing-with-a-recipe","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors > Defining the Workflow","what":"1. Preprocessing with a Recipe","title":"AMR with tidymodels","text":"create recipe preprocess data modelling. recipe includes least one preprocessing operation, like step_corr(), necessary parameters can estimated training set using prep(): Explanation: recipe(mo ~ ., data = data) take mo column outcome columns predictors. step_corr() removes predictors (.e., antibiotic columns) higher correlation 90%. Notice recipe contains just antimicrobial selector functions - need define columns specifically. preparation (retrieved prep()) can see columns variables ‘AMX’ ‘CTX’ removed correlate much existing, variables.","code":"# Define the recipe for data preprocessing resistance_recipe <- recipe(mo ~ ., data = data) %>% step_corr(c(aminoglycosides(), betalactams()), threshold = 0.9) resistance_recipe #> #> ── Recipe ────────────────────────────────────────────────────────────────────── #> #> ── Inputs #> Number of variables by role #> outcome: 1 #> predictor: 20 #> #> ── Operations #> • Correlation filter on: c(aminoglycosides(), betalactams()) prep(resistance_recipe) #> ℹ For `aminoglycosides()` using columns 'GEN' (gentamicin), 'TOB' #> (tobramycin), 'AMK' (amikacin), and 'KAN' (kanamycin) #> ℹ For `betalactams()` using columns 'PEN' (benzylpenicillin), 'OXA' #> (oxacillin), 'FLC' (flucloxacillin), 'AMX' (amoxicillin), 'AMC' #> (amoxicillin/clavulanic acid), 'AMP' (ampicillin), 'TZP' #> (piperacillin/tazobactam), 'CZO' (cefazolin), 'FEP' (cefepime), 'CXM' #> (cefuroxime), 'FOX' (cefoxitin), 'CTX' (cefotaxime), 'CAZ' (ceftazidime), #> 'CRO' (ceftriaxone), 'IPM' (imipenem), and 'MEM' (meropenem) #> #> ── Recipe ────────────────────────────────────────────────────────────────────── #> #> ── Inputs #> Number of variables by role #> outcome: 1 #> predictor: 20 #> #> ── Training information #> Training data contained 1968 data points and no incomplete rows. #> #> ── Operations #> • Correlation filter on: AMX CTX | Trained"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"specifying-the-model","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors > Defining the Workflow","what":"2. Specifying the Model","title":"AMR with tidymodels","text":"define logistic regression model since resistance prediction binary classification task. Explanation: logistic_reg() sets logistic regression model. set_engine(\"glm\") specifies use R’s built-GLM engine.","code":"# Specify a logistic regression model logistic_model <- logistic_reg() %>% set_engine(\"glm\") # Use the Generalised Linear Model engine logistic_model #> Logistic Regression Model Specification (classification) #> #> Computational engine: glm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"building-the-workflow","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors > Defining the Workflow","what":"3. Building the Workflow","title":"AMR with tidymodels","text":"bundle recipe model together workflow, organises entire modelling process.","code":"# Combine the recipe and model into a workflow resistance_workflow <- workflow() %>% add_recipe(resistance_recipe) %>% # Add the preprocessing recipe add_model(logistic_model) # Add the logistic regression model resistance_workflow #> ══ Workflow ════════════════════════════════════════════════════════════════════ #> Preprocessor: Recipe #> Model: logistic_reg() #> #> ── Preprocessor ──────────────────────────────────────────────────────────────── #> 1 Recipe Step #> #> • step_corr() #> #> ── Model ─────────────────────────────────────────────────────────────────────── #> Logistic Regression Model Specification (classification) #> #> Computational engine: glm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"training-and-evaluating-the-model","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors","what":"Training and Evaluating the Model","title":"AMR with tidymodels","text":"train model, split data training testing sets. , fit workflow training set evaluate performance. Explanation: initial_split() splits data training testing sets. fit() trains workflow training set. Notice fit(), antimicrobial selector functions internally called . training, functions called since stored recipe. Next, evaluate model testing data. Explanation: predict() generates predictions testing set. metrics() computes evaluation metrics like accuracy kappa. appears can predict Gram stain 99.5% accuracy based AMR results aminoglycosides beta-lactam antibiotics. ROC curve looks like :","code":"# Split data into training and testing sets set.seed(123) # For reproducibility data_split <- initial_split(data, prop = 0.8) # 80% training, 20% testing training_data <- training(data_split) # Training set testing_data <- testing(data_split) # Testing set # Fit the workflow to the training data fitted_workflow <- resistance_workflow %>% fit(training_data) # Train the model # Make predictions on the testing set predictions <- fitted_workflow %>% predict(testing_data) # Generate predictions probabilities <- fitted_workflow %>% predict(testing_data, type = \"prob\") # Generate probabilities predictions <- predictions %>% bind_cols(probabilities) %>% bind_cols(testing_data) # Combine with true labels predictions #> # A tibble: 394 × 24 #> .pred_class `.pred_Gram-negative` `.pred_Gram-positive` mo GEN TOB #> #> 1 Gram-positive 1.07e- 1 8.93 e- 1 Gram-p… 5 5 #> 2 Gram-positive 3.17e- 8 1.000e+ 0 Gram-p… 5 1 #> 3 Gram-negative 9.99e- 1 1.42 e- 3 Gram-n… 5 5 #> 4 Gram-positive 2.22e-16 1 e+ 0 Gram-p… 5 5 #> 5 Gram-negative 9.46e- 1 5.42 e- 2 Gram-n… 5 5 #> 6 Gram-positive 1.07e- 1 8.93 e- 1 Gram-p… 5 5 #> 7 Gram-positive 2.22e-16 1 e+ 0 Gram-p… 1 5 #> 8 Gram-positive 2.22e-16 1 e+ 0 Gram-p… 4 4 #> 9 Gram-negative 1 e+ 0 2.22 e-16 Gram-n… 1 1 #> 10 Gram-positive 6.05e-11 1.000e+ 0 Gram-p… 4 4 #> # ℹ 384 more rows #> # ℹ 18 more variables: AMK , KAN , PEN , OXA , FLC , #> # AMX , AMC , AMP , TZP , CZO , FEP , #> # CXM , FOX , CTX , CAZ , CRO , IPM , MEM # Evaluate model performance metrics <- predictions %>% metrics(truth = mo, estimate = .pred_class) # Calculate performance metrics metrics #> # A tibble: 2 × 3 #> .metric .estimator .estimate #> #> 1 accuracy binary 0.995 #> 2 kap binary 0.989 # To assess some other model properties, you can make our own `metrics()` function our_metrics <- metric_set(accuracy, kap, ppv, npv) # add Positive Predictive Value and Negative Predictive Value metrics2 <- predictions %>% our_metrics(truth = mo, estimate = .pred_class) # run again on our `our_metrics()` function metrics2 #> # A tibble: 4 × 3 #> .metric .estimator .estimate #> #> 1 accuracy binary 0.995 #> 2 kap binary 0.989 #> 3 ppv binary 0.987 #> 4 npv binary 1 predictions %>% roc_curve(mo, `.pred_Gram-negative`) %>% autoplot()"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"conclusion","dir":"Articles","previous_headings":"Example 1: Using Antimicrobial Selectors","what":"Conclusion","title":"AMR with tidymodels","text":"post, demonstrated build machine learning pipeline tidymodels framework AMR package. combining selector functions like aminoglycosides() betalactams() tidymodels, efficiently prepared data, trained model, evaluated performance. workflow extensible antimicrobial classes resistance patterns, empowering users analyse AMR data systematically reproducibly.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"example-2-predicting-esbl-presence-using-raw-mics","dir":"Articles","previous_headings":"","what":"Example 2: Predicting ESBL Presence Using Raw MICs","title":"AMR with tidymodels","text":"second example, demonstrate use columns directly tidymodels workflows using AMR-specific recipe steps. includes transformation log2 scale using step_mic_log2(), prepares MIC values use classification models. approach idea formed basis publication DOI: 10.3389/fmicb.2025.1582703 model presence extended-spectrum beta-lactamases (ESBL).","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"objective-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs","what":"Objective","title":"AMR with tidymodels","text":"goal : Use raw MIC values predict whether bacterial isolate produces ESBL. Apply AMR-aware preprocessing tidymodels recipe. Train classification model evaluate predictive performance.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"data-preparation-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs","what":"Data Preparation","title":"AMR with tidymodels","text":"use esbl_isolates dataset comes AMR package. Explanation: esbl_isolates: Contains MIC test results ESBL status isolate. mutate(esbl = ...): Converts target column ordered factor classification.","code":"# Load required libraries library(AMR) library(tidymodels) # View the esbl_isolates data set esbl_isolates #> # A tibble: 500 × 19 #> esbl genus AMC AMP TZP CXM FOX CTX CAZ GEN TOB TMP SXT #> #> 1 FALSE Esch… 32 32 4 64 64 8.00 8.00 1 1 16.0 20 #> 2 FALSE Esch… 32 32 4 64 64 4.00 8.00 1 1 16.0 320 #> 3 FALSE Esch… 4 2 64 8 4 8.00 0.12 16 16 0.5 20 #> 4 FALSE Kleb… 32 32 16 64 64 8.00 8.00 1 1 0.5 20 #> 5 FALSE Esch… 32 32 4 4 4 0.25 2.00 1 1 16.0 320 #> 6 FALSE Citr… 32 32 16 64 64 64.00 32.00 1 1 0.5 20 #> 7 FALSE Morg… 32 32 4 64 64 16.00 2.00 1 1 0.5 20 #> 8 FALSE Prot… 16 32 4 1 4 8.00 0.12 1 1 16.0 320 #> 9 FALSE Ente… 32 32 8 64 64 32.00 4.00 1 1 0.5 20 #> 10 FALSE Citr… 32 32 32 64 64 8.00 64.00 1 1 16.0 320 #> # ℹ 490 more rows #> # ℹ 6 more variables: NIT , FOS , CIP , IPM , MEM , #> # COL # Prepare a binary outcome and convert to ordered factor data <- esbl_isolates %>% mutate(esbl = factor(esbl, levels = c(FALSE, TRUE), ordered = TRUE))"},{"path":[]},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"preprocessing-with-a-recipe-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs > Defining the Workflow","what":"1. Preprocessing with a Recipe","title":"AMR with tidymodels","text":"use step_mic_log2() function log2-transform MIC values, ensuring MICs numeric properly scaled. MIC predictors can easily agnostically selected using new all_mic_predictors(): Explanation: remove_role(): Removes irrelevant variables like genus. step_mic_log2(): Applies log2(.numeric(...)) MIC predictors one go. prep(): Finalises recipe based training data.","code":"# Split into training and testing sets set.seed(123) split <- initial_split(data) training_data <- training(split) testing_data <- testing(split) # Define the recipe mic_recipe <- recipe(esbl ~ ., data = training_data) %>% remove_role(genus, old_role = \"predictor\") %>% # Remove non-informative variable step_mic_log2(all_mic_predictors()) #%>% # Log2 transform all MIC predictors # prep() mic_recipe #> #> ── Recipe ────────────────────────────────────────────────────────────────────── #> #> ── Inputs #> Number of variables by role #> outcome: 1 #> predictor: 17 #> undeclared role: 1 #> #> ── Operations #> • Log2 transformation of MIC columns: all_mic_predictors()"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"specifying-the-model-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs > Defining the Workflow","what":"2. Specifying the Model","title":"AMR with tidymodels","text":"use simple logistic regression model ESBL presence, though recent models xgboost (link parsnip manual) much precise. Explanation: logistic_reg(): Specifies binary classification model. set_engine(\"glm\"): Uses base R GLM engine.","code":"# Define the model model <- logistic_reg(mode = \"classification\") %>% set_engine(\"glm\") model #> Logistic Regression Model Specification (classification) #> #> Computational engine: glm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"building-the-workflow-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs > Defining the Workflow","what":"3. Building the Workflow","title":"AMR with tidymodels","text":"","code":"# Create workflow workflow_model <- workflow() %>% add_recipe(mic_recipe) %>% add_model(model) workflow_model #> ══ Workflow ════════════════════════════════════════════════════════════════════ #> Preprocessor: Recipe #> Model: logistic_reg() #> #> ── Preprocessor ──────────────────────────────────────────────────────────────── #> 1 Recipe Step #> #> • step_mic_log2() #> #> ── Model ─────────────────────────────────────────────────────────────────────── #> Logistic Regression Model Specification (classification) #> #> Computational engine: glm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"training-and-evaluating-the-model-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs","what":"Training and Evaluating the Model","title":"AMR with tidymodels","text":"Explanation: fit(): Trains model processed training data. predict(): Produces predictions unseen test data. metric_set(): Allows evaluating multiple classification metrics. appears can predict ESBL gene presence positive predictive value (PPV) 92.1% negative predictive value (NPV) 91.9 using simplistic logistic regression model.","code":"# Fit the model fitted <- fit(workflow_model, training_data) # Generate predictions predictions <- predict(fitted, testing_data) %>% bind_cols(testing_data) # Evaluate model performance our_metrics <- metric_set(accuracy, kap, ppv, npv) metrics <- our_metrics(predictions, truth = esbl, estimate = .pred_class) metrics #> # A tibble: 4 × 3 #> .metric .estimator .estimate #> #> 1 accuracy binary 0.92 #> 2 kap binary 0.840 #> 3 ppv binary 0.921 #> 4 npv binary 0.919"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"visualising-predictions","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs","what":"Visualising Predictions","title":"AMR with tidymodels","text":"can visualise predictions comparing predicted actual ESBL status.","code":"library(ggplot2) ggplot(predictions, aes(x = esbl, fill = .pred_class)) + geom_bar(position = \"stack\") + labs(title = \"Predicted vs Actual ESBL Status\", x = \"Actual ESBL\", y = \"Count\") + theme_minimal()"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"conclusion-1","dir":"Articles","previous_headings":"Example 2: Predicting ESBL Presence Using Raw MICs","what":"Conclusion","title":"AMR with tidymodels","text":"example, showcased new AMR-specific recipe steps simplify working columns tidymodels. step_mic_log2() transformation converts ordered MICs log2-transformed numerics, improving compatibility classification models. pipeline enables realistic, reproducible, interpretable modelling antimicrobial resistance data.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"example-3-predicting-amr-over-time","dir":"Articles","previous_headings":"","what":"Example 3: Predicting AMR Over Time","title":"AMR with tidymodels","text":"third example, aim predict antimicrobial resistance (AMR) trends time using tidymodels. model resistance three antibiotics (amoxicillin AMX, amoxicillin-clavulanic acid AMC, ciprofloxacin CIP), based historical data grouped year hospital ward.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"objective-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Objective","title":"AMR with tidymodels","text":"goal : Prepare dataset aggregating resistance data time. Define regression model predict AMR trends. Use tidymodels preprocess, train, evaluate model.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"data-preparation-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Data Preparation","title":"AMR with tidymodels","text":"start transforming example_isolates dataset structured time-series format. Explanation: mo_name(mo): Converts microbial codes proper species names. resistance(): Converts AMR results numeric values (proportion resistant isolates). group_by(year, ward, species): Aggregates resistance rates year ward.","code":"# Load required libraries library(AMR) library(tidymodels) # Transform dataset data_time <- example_isolates %>% top_n_microorganisms(n = 10) %>% # Filter on the top #10 species mutate(year = as.integer(format(date, \"%Y\")), # Extract year from date gramstain = mo_gramstain(mo)) %>% # Get taxonomic names group_by(year, gramstain) %>% summarise(across(c(AMX, AMC, CIP), function(x) resistance(x, minimum = 0), .names = \"res_{.col}\"), .groups = \"drop\") %>% filter(!is.na(res_AMX) & !is.na(res_AMC) & !is.na(res_CIP)) # Drop missing values #> ℹ Using column 'mo' as input for `col_mo`. data_time #> # A tibble: 32 × 5 #> year gramstain res_AMX res_AMC res_CIP #> #> 1 2002 Gram-negative 1 0.105 0.0606 #> 2 2002 Gram-positive 0.838 0.182 0.162 #> 3 2003 Gram-negative 1 0.0714 0 #> 4 2003 Gram-positive 0.714 0.244 0.154 #> 5 2004 Gram-negative 0.464 0.0938 0 #> 6 2004 Gram-positive 0.849 0.299 0.244 #> 7 2005 Gram-negative 0.412 0.132 0.0588 #> 8 2005 Gram-positive 0.882 0.382 0.154 #> 9 2006 Gram-negative 0.379 0 0.1 #> 10 2006 Gram-positive 0.778 0.333 0.353 #> # ℹ 22 more rows"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"defining-the-workflow-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Defining the Workflow","title":"AMR with tidymodels","text":"now define modelling workflow, consists preprocessing step, model specification, fitting process.","code":""},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"preprocessing-with-a-recipe-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time > Defining the Workflow","what":"1. Preprocessing with a Recipe","title":"AMR with tidymodels","text":"Explanation: step_dummy(): Encodes categorical variables (ward, species) numerical indicators. step_normalize(): Normalises year variable. step_nzv(): Removes near-zero variance predictors.","code":"# Define the recipe resistance_recipe_time <- recipe(res_AMX ~ year + gramstain, data = data_time) %>% step_dummy(gramstain, one_hot = TRUE) %>% # Convert categorical to numerical step_normalize(year) %>% # Normalise year for better model performance step_nzv(all_predictors()) # Remove near-zero variance predictors resistance_recipe_time #> #> ── Recipe ────────────────────────────────────────────────────────────────────── #> #> ── Inputs #> Number of variables by role #> outcome: 1 #> predictor: 2 #> #> ── Operations #> • Dummy variables from: gramstain #> • Centering and scaling for: year #> • Sparse, unbalanced variable filter on: all_predictors()"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"specifying-the-model-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time > Defining the Workflow","what":"2. Specifying the Model","title":"AMR with tidymodels","text":"use linear regression model predict resistance trends. Explanation: linear_reg(): Defines linear regression model. set_engine(\"lm\"): Uses R’s built-linear regression engine.","code":"# Define the linear regression model lm_model <- linear_reg() %>% set_engine(\"lm\") # Use linear regression lm_model #> Linear Regression Model Specification (regression) #> #> Computational engine: lm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"building-the-workflow-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time > Defining the Workflow","what":"3. Building the Workflow","title":"AMR with tidymodels","text":"combine preprocessing recipe model workflow.","code":"# Create workflow resistance_workflow_time <- workflow() %>% add_recipe(resistance_recipe_time) %>% add_model(lm_model) resistance_workflow_time #> ══ Workflow ════════════════════════════════════════════════════════════════════ #> Preprocessor: Recipe #> Model: linear_reg() #> #> ── Preprocessor ──────────────────────────────────────────────────────────────── #> 3 Recipe Steps #> #> • step_dummy() #> • step_normalize() #> • step_nzv() #> #> ── Model ─────────────────────────────────────────────────────────────────────── #> Linear Regression Model Specification (regression) #> #> Computational engine: lm"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"training-and-evaluating-the-model-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Training and Evaluating the Model","title":"AMR with tidymodels","text":"split data training testing sets, fit model, evaluate performance. Explanation: initial_split(): Splits data training testing sets. fit(): Trains workflow. predict(): Generates resistance predictions. metrics(): Evaluates model performance.","code":"# Split the data set.seed(123) data_split_time <- initial_split(data_time, prop = 0.8) train_time <- training(data_split_time) test_time <- testing(data_split_time) # Train the model fitted_workflow_time <- resistance_workflow_time %>% fit(train_time) # Make predictions predictions_time <- fitted_workflow_time %>% predict(test_time) %>% bind_cols(test_time) # Evaluate model metrics_time <- predictions_time %>% metrics(truth = res_AMX, estimate = .pred) metrics_time #> # A tibble: 3 × 3 #> .metric .estimator .estimate #> #> 1 rmse standard 0.0774 #> 2 rsq standard 0.711 #> 3 mae standard 0.0704"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"visualising-predictions-1","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Visualising Predictions","title":"AMR with tidymodels","text":"plot resistance trends time amoxicillin. Additionally, can visualise resistance trends ggplot2 directly add linear models :","code":"library(ggplot2) # Plot actual vs predicted resistance over time ggplot(predictions_time, aes(x = year)) + geom_point(aes(y = res_AMX, color = \"Actual\")) + geom_line(aes(y = .pred, color = \"Predicted\")) + labs(title = \"Predicted vs Actual AMX Resistance Over Time\", x = \"Year\", y = \"Resistance Proportion\") + theme_minimal() ggplot(data_time, aes(x = year, y = res_AMX, color = gramstain)) + geom_line() + labs(title = \"AMX Resistance Trends\", x = \"Year\", y = \"Resistance Proportion\") + # add a linear model directly in ggplot2: geom_smooth(method = \"lm\", formula = y ~ x, alpha = 0.25) + theme_minimal()"},{"path":"https://amr-for-r.org/articles/AMR_with_tidymodels.html","id":"conclusion-2","dir":"Articles","previous_headings":"Example 3: Predicting AMR Over Time","what":"Conclusion","title":"AMR with tidymodels","text":"example, demonstrated analyze AMR trends time using tidymodels. aggregating resistance rates year hospital ward, built predictive model track changes resistance amoxicillin (AMX), amoxicillin-clavulanic acid (AMC), ciprofloxacin (CIP). method can extended antibiotics resistance patterns, providing valuable insights AMR dynamics healthcare settings.","code":""},{"path":"https://amr-for-r.org/articles/EUCAST.html","id":"introduction","dir":"Articles","previous_headings":"","what":"Introduction","title":"Apply EUCAST rules","text":"EUCAST rules? European Committee Antimicrobial Susceptibility Testing (EUCAST) states website: EUCAST expert rules (see ) tabulated collection expert knowledge interpretive rules, expected resistant phenotypes expected susceptible phenotypes applied antimicrobial susceptibility testing order reduce testing, reduce errors make appropriate recommendations reporting particular resistances. Europe, lot medical microbiological laboratories already apply rules (Brown et al., 2015). package features latest insights expected resistant phenotypes (v1.2, 2023).","code":""},{"path":"https://amr-for-r.org/articles/EUCAST.html","id":"examples","dir":"Articles","previous_headings":"","what":"Examples","title":"Apply EUCAST rules","text":"rules can used discard improbable bug-drug combinations data. example, Klebsiella produces beta-lactamase prevents ampicillin (amoxicillin) working . words, practically every strain Klebsiella resistant ampicillin. Sometimes, laboratory data can still contain strains Klebsiella susceptible ampicillin. antibiogram available identification available, antibiogram re-interpreted based identification. eucast_rules() function resolves , applying latest ‘EUCAST Expected Resistant Phenotypes’ guideline: convenient function mo_is_intrinsic_resistant() uses guideline, allows check one specific microorganisms antimicrobials: EUCAST rules can used correction, can also used filling known resistance susceptibility based results antimicrobials drugs. process called interpretive reading, basically form imputation:","code":"oops <- tibble::tibble( mo = c( \"Klebsiella pneumoniae\", \"Escherichia coli\" ), ampicillin = as.sir(\"S\") ) oops #> # A tibble: 2 × 2 #> mo ampicillin #> #> 1 Klebsiella pneumoniae S #> 2 Escherichia coli S eucast_rules(oops, info = FALSE, overwrite = TRUE) #> # A tibble: 2 × 2 #> mo ampicillin #> #> 1 Klebsiella pneumoniae R #> 2 Escherichia coli S mo_is_intrinsic_resistant( c(\"Klebsiella pneumoniae\", \"Escherichia coli\"), \"ampicillin\" ) #> [1] TRUE FALSE mo_is_intrinsic_resistant( \"Klebsiella pneumoniae\", c(\"ampicillin\", \"kanamycin\") ) #> [1] TRUE FALSE data <- tibble::tibble( mo = c( \"Staphylococcus aureus\", \"Enterococcus faecalis\", \"Escherichia coli\", \"Klebsiella pneumoniae\", \"Pseudomonas aeruginosa\" ), VAN = \"-\", # Vancomycin AMX = \"-\", # Amoxicillin COL = \"-\", # Colistin CAZ = \"-\", # Ceftazidime CXM = \"-\", # Cefuroxime PEN = \"S\", # Benzylenicillin FOX = \"S\" # Cefoxitin ) data eucast_rules(data, overwrite = TRUE)"},{"path":[]},{"path":"https://amr-for-r.org/articles/PCA.html","id":"transforming","dir":"Articles","previous_headings":"","what":"Transforming","title":"Conduct principal component analysis (PCA) for AMR","text":"PCA, need transform AMR data first. example_isolates data set package looks like: Now transform data set resistance percentages per taxonomic order genus:","code":"library(AMR) library(dplyr) glimpse(example_isolates) #> Rows: 2,000 #> Columns: 46 #> $ date 2002-01-02, 2002-01-03, 2002-01-07, 2002-01-07, 2002-01-13, 2… #> $ patient \"A77334\", \"A77334\", \"067927\", \"067927\", \"067927\", \"067927\", \"4… #> $ age 65, 65, 45, 45, 45, 45, 78, 78, 45, 79, 67, 67, 71, 71, 75, 50… #> $ gender \"F\", \"F\", \"F\", \"F\", \"F\", \"F\", \"M\", \"M\", \"F\", \"F\", \"M\", \"M\", \"M… #> $ ward \"Clinical\", \"Clinical\", \"ICU\", \"ICU\", \"ICU\", \"ICU\", \"Clinical\"… #> $ mo \"B_ESCHR_COLI\", \"B_ESCHR_COLI\", \"B_STPHY_EPDR\", \"B_STPHY_EPDR\",… #> $ PEN R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, R, S,… #> $ OXA NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA… #> $ FLC NA, NA, R, R, R, R, S, S, R, S, S, S, NA, NA, NA, NA, NA, R, R… #> $ AMX NA, NA, NA, NA, NA, NA, R, R, NA, NA, NA, NA, NA, NA, R, NA, N… #> $ AMC I, I, NA, NA, NA, NA, S, S, NA, NA, S, S, I, I, R, I, I, NA, N… #> $ AMP NA, NA, NA, NA, NA, NA, R, R, NA, NA, NA, NA, NA, NA, R, NA, N… #> $ TZP 