(v0.8.0.9002) eucast_rules() fix for S. maltophilia

data-raw/eucast_rules.tsv

@@ -137,7 +137,7 @@ fullname	like	^Burkholderia (cepacia|multivorans|cenocepacia|stabilis|vietnamien
 genus_species	is	Elizabethkingia meningoseptica			aminopenicillins, AMC, TIC, CZO, CTX, CRO, CAZ, FEP, ATM, ETP, IPM, MEM, polymyxins	R	Table 02: Intrinsic resistance in non-fermentative Gram-negative bacteria	Expert Rules
 genus_species	is	Ochrobactrum anthropi			aminopenicillins, AMC, TIC, PIP, TZP, CZO, CTX, CRO, CAZ, FEP, ATM, ETP	R	Table 02: Intrinsic resistance in non-fermentative Gram-negative bacteria	Expert Rules
 genus_species	is	Pseudomonas aeruginosa			aminopenicillins, AMC, CZO, CTX, CRO, ETP, CHL, KAN, NEO, TMP, SXT, tetracyclines, TGC	R	Table 02: Intrinsic resistance in non-fermentative Gram-negative bacteria	Expert Rules
-genus_species	is	Stenotrophomonas maltophilia			aminopenicillins, AMC, TIC, PIP, TZP, CZO, CTX, CRO, CAZ, ATM, ETP, IPM, MEM, aminoglycosides, TMP, FOS, TCY	R	Table 02: Intrinsic resistance in non-fermentative Gram-negative bacteria	Expert Rules
+genus_species	is	Stenotrophomonas maltophilia			aminopenicillins, AMC, TIC, PIP, TZP, CZO, CTX, CRO, ATM, ETP, IPM, MEM, aminoglycosides, TMP, FOS, TCY	R	Table 02: Intrinsic resistance in non-fermentative Gram-negative bacteria	Expert Rules
 genus	one_of	Haemophilus, Moraxella, Neisseria, Campylobacter			glycopeptides, LIN, DAP, LNZ	R	Table 03: Intrinsic resistance in other Gram-negative bacteria	Expert Rules
 genus_species	is	Haemophilus influenzae			FUS, streptogramins	R	Table 03: Intrinsic resistance in other Gram-negative bacteria	Expert Rules
 genus_species	is	Moraxella catarrhalis			TMP	R	Table 03: Intrinsic resistance in other Gram-negative bacteria	Expert Rules

data-raw/internals.R

@@ -1,13 +1,7 @@
-# ---------------------------------------------------------------------------------------------------
-# For editing this EUCAST reference file, these values can all be used for target antibiotics:
-# all_betalactams, aminoglycosides, carbapenems, cephalosporins, cephalosporins_without_CAZ, fluoroquinolones, 
-# glycopeptides, macrolides, minopenicillins, polymyxins, streptogramins, tetracyclines, ureidopenicillins
-# and all separate EARS-Net letter codes like AMC. They can be separated by comma: 'AMC, fluoroquinolones'.
-# The if_mo_property column can be any column name from the AMR::microorganisms data set, or "genus_species" or "gramstain".
-# The EUCAST guideline contains references to the 'Burkholderia cepacia complex'. All species in this group can be found in:
-# LiPuma J, Curr Opin Pulm Med. 2005 Nov;11(6):528-33. (PMID 16217180).
-# >>>>> IF YOU WANT TO IMPORT THIS FILE INTO YOUR OWN SOFTWARE, HAVE THE FIRST 10 LINES SKIPPED <<<<<
-# ---------------------------------------------------------------------------------------------------
+# Run this file to update the package -------------------------------------
+# source("data-raw/internals.R")
+
+# See 'data-raw/eucast_rules.tsv' for the EUCAST reference file
 eucast_rules_file <- dplyr::arrange(
   .data = utils::read.delim(file = "data-raw/eucast_rules.tsv",
                             skip = 10,
@@ -19,6 +13,7 @@ eucast_rules_file <- dplyr::arrange(
   reference.rule_group,
   reference.rule)
 
+
 # Translations ----
 translations_file <- utils::read.delim(file = "data-raw/translations.tsv",
                                        sep = "\t",

data-raw/reproduction_of_microorganisms.R

@@ -70,7 +70,13 @@ data_dsmz <- data_dsmz %>%
 # DSMZ only contains genus/(sub)species, try to find taxonomic properties based on genus and data_col
 ref_taxonomy <- data_col %>%
   filter(genus %in% data_dsmz$genus,
+         kingdom %in% c("Bacteria", "Chromista", "Archaea", "Protozoa", "Fungi"),
          family != "") %>%
+  mutate(kingdom = factor(kingdom,
+                          # in the left_join following, try Bacteria first, then Chromista, ...
+                          levels = c("Bacteria", "Chromista", "Archaea", "Protozoa", "Fungi"),
+                          ordered = TRUE)) %>% 
+  arrange(kingdom) %>% 
   distinct(genus, .keep_all = TRUE) %>%
   select(kingdom, phylum, class, order, family, genus)
 
@@ -197,6 +203,7 @@ MOs <- MOs %>%
 MOs$ref[!grepl("^d[A-Z]", MOs$ref)] <- gsub("^([a-z])", "\\U\\1", MOs$ref[!grepl("^d[A-Z]", MOs$ref)], perl = TRUE)
 # specific one for the French that are named dOrbigny 
 MOs$ref[grepl("^d[A-Z]", MOs$ref)] <- gsub("^d", "d'", MOs$ref[grepl("^d[A-Z]", MOs$ref)])
+MOs <- MOs %>% mutate(ref = gsub(" +", " ", ref))
 
 # Remove non-ASCII characters (these are not allowed by CRAN)
 MOs <- MOs %>%
@@ -275,9 +282,15 @@ MOs <- MOs %>%
             by = "kingdom_fullname",
             suffix = c("_dsmz", "_col")) %>% 
   mutate(col_id = col_id_col, 
-         species_id = ifelse(!is.na(species_id_col), gsub(".*/(.*)$", "\\1", species_id_col), species_id_dsmz),
-         source = ifelse(!is.na(species_id_col), source_col, source_dsmz), 
-         ref = ifelse(!is.na(species_id_col) & ref_col != "", ref_col, ref_dsmz)) %>% 
+         species_id = ifelse(!is.na(species_id_col) & ref_col == ref_dsmz, 
+                             gsub(".*/(.*)$", "\\1", species_id_col), 
+                             species_id_dsmz),
+         source = ifelse(!is.na(species_id_col) & ref_col == ref_dsmz,
+                         source_col, 
+                         source_dsmz), 
+         ref = ifelse(!is.na(species_id_col) & ref_col == ref_dsmz,
+                      ref_col, 
+                      ref_dsmz)) %>% 
   select(-matches("(_col|_dsmz|kingdom_fullname)"))
 
 
@@ -296,6 +309,7 @@ sum(MOs.old$fullname %in% MOs$fullname)
 
 # what characters are in the fullnames?
 table(sort(unlist(strsplit(x = paste(MOs$fullname, collapse = ""), split = ""))))
+MOs %>% filter(!fullname %like% "^[a-z ]+$") %>% View()
 
 table(MOs$kingdom, MOs$rank)
 table(AMR::microorganisms$kingdom, AMR::microorganisms$rank)
@@ -676,8 +690,16 @@ old_new <- MOs %>%
   left_join(AMR::microorganisms %>% mutate(kingdom_fullname = paste(kingdom, fullname)) %>% select(mo, kingdom_fullname), by = "kingdom_fullname", suffix = c("_new", "_old")) %>% 
   filter(mo_new != mo_old) %>% 
   select(mo_old, mo_new, everything())
-old_new %>% 
-  View()
+
+View(old_new)
+# to keep all the old IDs:
+# MOs <- MOs %>% filter(!mo %in% old_new$mo_new) %>% 
+#   rbind(microorganisms %>%
+#           filter(mo %in% old_new$mo_old) %>%
+#           select(mo, fullname) %>%
+#           left_join(MOs %>% 
+#                       select(-mo), by = "fullname"))
+
 # and these codes are now missing (which will throw a unit test error):
 AMR::microorganisms.codes %>% filter(!mo %in% MOs$mo)
 AMR::rsi_translation %>% filter(!mo %in% MOs$mo)

data-raw/reproduction_of_rsi_translation.R

@@ -1,16 +1,16 @@
 library(dplyr)
-library(readxl)
 
 # Installed WHONET 2019 software on Windows (http://www.whonet.org/software.html),
 #    opened C:\WHONET\Codes\WHONETCodes.mdb in MS Access
 #    and exported table 'DRGLST1' to MS Excel
-DRGLST1 <- read_excel("data-raw/DRGLST1.xlsx")
+DRGLST1 <- readxl::read_excel("data-raw/DRGLST1.xlsx")
 rsi_translation <- DRGLST1 %>%
   # only keep CLSI and EUCAST guidelines:
   filter(GUIDELINES %like% "^(CLSI|EUCST)") %>%
   # set a nice layout:
   transmute(guideline = gsub("([0-9]+)$", " 20\\1", gsub("EUCST", "EUCAST", GUIDELINES)),
             method = TESTMETHOD,
+            site = SITE_INF,
             mo = as.mo(ORG_CODE),
             ab = as.ab(WHON5_CODE),
             ref_tbl = REF_TABLE,
@@ -19,7 +19,7 @@ rsi_translation <- DRGLST1 %>%
             R_disk = as.disk(DISK_R),
             S_mic = as.mic(MIC_S),
             R_mic = as.mic(MIC_R)) %>%
-  filter(!is.na(mo) & !is.na(ab)) %>%
+  filter(!is.na(mo) & !is.na(ab) & !mo %in% c("UNKNOWN", "B_GRAMN", "B_GRAMP", "F_FUNGUS", "F_YEAST")) %>%
   arrange(desc(guideline), mo, ab)
 
 print(mo_failures())
@@ -27,27 +27,20 @@ print(mo_failures())
 # create 2 tables: MIC and disk
 tbl_mic <- rsi_translation %>%
   filter(method == "MIC") %>%
-  select(-ends_with("_disk")) %>%
-  mutate(joinstring = paste(guideline, mo, ab))
+  mutate(breakpoint_S = as.double(S_mic), breakpoint_R = as.double(R_mic))
 tbl_disk <- rsi_translation %>%
   filter(method == "DISK") %>%
-  select(-S_mic, -R_mic) %>%
-  mutate(joinstring = paste(guideline, mo, ab)) %>%
-  select(joinstring, ends_with("_disk"))
+  mutate(breakpoint_S = as.double(S_disk), breakpoint_R = as.double(R_disk))
 
 # merge them so every record is a unique combination of method, mo and ab
-rsi_translation <- tbl_mic %>%
-  left_join(tbl_disk,
-            by = "joinstring") %>%
-  select(-joinstring, -method) %>%
+rsi_translation <- bind_rows(tbl_mic, tbl_disk) %>%
+  rename(disk_dose = dose_disk) %>% 
+  mutate(disk_dose = gsub("µ", "u", disk_dose)) %>% 
+  select(-ends_with("_mic"), -ends_with("_disk")) %>%
   as.data.frame(stringsAsFactors = FALSE) %>%
   # force classes again
   mutate(mo = as.mo(mo),
-         ab = as.ab(ab),
-         S_mic = as.mic(S_mic),
-         R_mic = as.mic(R_mic),
-         S_disk = as.disk(S_disk),
-         R_disk = as.disk(R_disk))
+         ab = as.ab(ab))
 
 # save to package
 usethis::use_data(rsi_translation, overwrite = TRUE)