In Zhenglei-BCS/beemixtox_public: Reproduce the manuscript work
knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>"
)
library(beemixtox1)
library(tidyverse)
library(knitr)
The purpose of this analysis is to make sure the data used in the manuscript is comparable to data from other databases. (Norman EMPODAT database, SUSDAT, EHype v3, SSD data from Posthuma)
- Data selection and curation:
- Unit with UGB UGB = microgram/bee (We assume generally all data is reported as ug ai/bee unless TOXDATA include reference to “form” or “product”)
- Remove all entries reported as ppb or ppm (these are not typical study designs)
- Limit to chemicals with at least 4 LD50 reported
- Exclude the following entries in the TOXDATA:
- 0.18%
- Low
- 27.7/57.4 ai
- 359 in air
- 359 in air
- 1083 in air
- 23 in vapor
- Exlclude the following entries with CHEMICAL
- Malathion Fyfanon 440g/L formulation
- Pyraclostrobin/Triticonazole BAS 673-01F mixture
- For those entries with TOXDATA including “form”, “product”, correct TOXDATA value by the “AI” column (which is the % active ingredient in the formulation)
unique(toxdata$DOSETYPE)
beetox_OPP <- toxdata %>% filter(COMMONNAME=="Honey bee" & DOSETYPE %in% c("LC50", "LD50") & TGL=="") %>% filter(TOXLEVEL %in% c("UGB")) %>% filter(STUDYTIME %in% c("48 hr","24 hr","3 D", "44 hr","48 H","48 Hr") ) %>% filter(!(CHEMICAL %in% c("Malathion Fyfanon 440g/L formulation","Pyraclostrobin/Triticonazole BAS 673-01F mixture")))
dim(beetox_OPP)
beetox_OPP <- beetox_OPP %>% filter(!(TOXICITY %in% c("0.18%", "Low", "27.7/57.4 ai", "359 in air", "1083 in air", "23 in vapor")))
id1 <- grep("/",beetox_OPP$AI)
beetox_OPP[id1,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Mixtures and Formulations, mixtures will be excluded; Replace g/L assuming density = 1. So, 100g/L = 10%, 200g/L = 20%, etc.")%>%kableExtra::kable_paper()
# [1] "95/5" "34/18" "56/6.2" "14/28" "200g/L" "100g/L" "20.6/20" "20.6/20" "1/7.5" "1/7.5" "13.3/5"
#[12] "0.5/0.1" "58/6" "6/58" "17/26" "1.4/2.1"
### Remove all "/" entries since they are mixtures.
## Replace g/L assuming density = 1. So, 100g/L = 10%, 200g/L = 20%, etc.
beetox_OPP <-beetox_OPP %>% mutate(AI=plyr::mapvalues(AI,from=c("200g/L","100g/L"),to=c("20","10")))
id1 <- grep("/",beetox_OPP$AI)
beetox_OPP <- beetox_OPP[-id1,]
grep("/",beetox_OPP$AI)
id <- c(grep("form",beetox_OPP$TOXICITY),grep("product",beetox_OPP$TOXICITY))
grep("/",beetox_OPP$AI[id])
### There is no form or product anymore!
id2 <- c(grep("WG",beetox_OPP$AI),grep("EC",beetox_OPP$AI),grep("SC",beetox_OPP$AI))
beetox_OPP[id2,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Formulations: Take the number as %")%>%kableExtra::kable_paper()
## For these ones remove the EC/SD or WG and use the nubmer as percentage?
beetox_OPP <- beetox_OPP %>% mutate(ai=gsub("WG","",AI)) %>% mutate(ai=gsub("SC","",ai)) %>% mutate(ai=gsub("EC","",ai))
## beetox_OPP[id2,"ai"]
## Now
Consider all the values that cannot be converted to a %.
id3 <-which(is.na(as.numeric(beetox_OPP$ai)))
unique(beetox_OPP$AI[id3])
beetox_OPP[id3,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Other non-numeric AI entries")%>%kableExtra::kable_paper()
For Tech grade make the ai % 100%.
Keep all "N.R" as 100%.
######
beetox_OPP$ai[grep("Tech",beetox_OPP$AI)] <- 100
beetox_OPP$ai[grep("tech",beetox_OPP$AI)] <- 100
beetox_OPP$ai[grep("N.R",beetox_OPP$AI)] <- 100
################
id4 <- beetox_OPP$AI %in% (unique(beetox_OPP$AI[id3])[-c(1,4)])
beetox_OPP[id4,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="N.R, FORM, ISOM: NR being kept as 100%, other s excluded")%>%kableExtra::kable_paper()
Exclude all other.
id5 <- unique(c(grep("FORM",beetox_OPP$AI),grep("Form",beetox_OPP$AI),grep("ISOM",beetox_OPP$AI)))
beetox_OPP <- beetox_OPP[-id5,] ### could be -id4
Check if ai is included in the Toxicity column
id6 <- unique(c(grep("ai",beetox_OPP$TOXICITY),grep("as",beetox_OPP$TOXICITY)))
beetox_OPP[id6,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Reported as AI or AS, but with AI entries not 100%, changed to 100%")%>%kableExtra::kable_paper()
beetox_OPP$ai[id6] <- 100
beetox_OPP$ai <- as.numeric(beetox_OPP$ai)
sum(is.na(beetox_OPP$ai))
## For those entries with TOXDATA including “form”, “product”, correct TOXDATA value by the “AI” column (which is the % active ingredient in the formulation)
beetox_OPP <- beetox_OPP %>% mutate(tox=as.numeric(unlist(lapply(str_split(TOXICITY," "),function(x)x[1]))))
summary(beetox_OPP$tox)
#beetox_OPP$tox[beetox_OPP$TOXLEVEL=="PPB"] <- beetox_OPP$tox[beetox_OPP$TOXLEVEL=="PPB"]/1000
#beetox_OPP$TOXLEVEL[beetox_OPP$TOXLEVEL=="PPB"] <- "PPM"
beetox_OPP$tox_ai <- beetox_OPP$tox*beetox_OPP$ai/100
sd(beetox_OPP$tox,na.rm = T) / mean(beetox_OPP$tox,na.rm = T)
## [1] 4.669772
beetox_OPP <- beetox_OPP %>% filter(!is.na(tox))
tmp <- beetox_OPP %>% group_by(CAS_NO)%>% summarise(ncas=n()) ## 159 CAS
beetox_OPP <- left_join(beetox_OPP,tmp)
beetox_OPP <- beetox_OPP%>% filter(ncas>3)
length(unique(beetox_OPP$CAS_NO)) ## 45 CAS ==> 23 cas
library(skimr)
cv <- function(x,na.rm=T){
sd(x,na.rm = na.rm)/mean(x,na.rm = na.rm)
}
my_skim <- skim_with(numeric = sfl(cv,nobs=length))
Not multiplied by AI column
cv_sum <- beetox_OPP %>% group_by(CAS_NO) %>% summarise(mean_LD=mean(tox),sd_LD=sd(tox),cv=sd_LD/mean_LD)
cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="Summary stats for overall CAS LD mean, sd, and CV ")%>%kableExtra::kable_classic()
beetox_OPP %>% left_join(casinfo_OPP,by=c("CAS_NO"="INPUT"))%>% group_by(CAS_NO,PREFERRED_NAME) %>% my_skim(tox) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table: Summary stats for LD mean, sd, and CV for each CAS</strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Multiplied by AI column
cv_sum <- beetox_OPP %>% group_by(CAS_NO) %>% summarise(mean_LD=mean(tox_ai),sd_LD=sd(tox_ai),cv=sd_LD/mean_LD)
cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="Summary stats for overall CAS LD mean, sd, and CV ")%>%kableExtra::kable_classic()
beetox_OPP %>% left_join(casinfo_OPP,by=c("CAS_NO"="INPUT"))%>% group_by(CAS_NO,PREFERRED_NAME) %>% my_skim(tox_ai) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table: Summary stats for LD mean, sd, and CV for each CAS</strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Ecotox DB Data
Assuming all ug/bee are ai ug/bee
Data.2$CAS.Number <- webchem::as.cas(Data.2$CAS.Number)
#Data.3 <- left_join(Data.2,Data.SSD%>%mutate(CAS.Number=as.character(CAS.)))
Data.3 <- left_join(Data.2,casinfo,by=c("CAS.Number"="INPUT"))
cv_sum <- Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% summarise(mean_LD=mean(Observed.Response.Mean),sd_LD=sd(Observed.Response.Mean),cv=sd_LD/mean_LD)
cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S1: Summary stats for overall CAS LD mean, sd, and CV </strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% my_skim(Observed.Response.Mean) %>% yank("numeric") %>% select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S2: Summary stats for LD mean, sd, and CV for each CAS</strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Use the purity column and consider values tranformed to AI ug/bee data
Data.3 <- Data.3 %>%filter(!is.na(Data.3$Observed.Response.Mean.t))
cv_sum <- Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% summarise(mean_LD.t=mean(Observed.Response.Mean.t),sd_LD.t=sd(Observed.Response.Mean.t),cv=sd_LD.t/mean_LD.t)
cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S1: Summary stats for overall CAS LD mean, sd, and CV </strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% my_skim(Observed.Response.Mean.t) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S2: Summary stats for LD mean, sd, and CV for each CAS</strong></center>",
escape = FALSE)%>%kableExtra::kable_classic()
Zhenglei-BCS/beemixtox_public documentation built on March 28, 2023, 1:51 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
knitr::opts_chunk$set( collapse = TRUE, comment = "#>" )
library(beemixtox1) library(tidyverse) library(knitr)
The purpose of this analysis is to make sure the data used in the manuscript is comparable to data from other databases. (Norman EMPODAT database, SUSDAT, EHype v3, SSD data from Posthuma)
- Data selection and curation:
- Unit with UGB UGB = microgram/bee (We assume generally all data is reported as ug ai/bee unless TOXDATA include reference to “form” or “product”)
- Remove all entries reported as ppb or ppm (these are not typical study designs)
- Limit to chemicals with at least 4 LD50 reported
- Exclude the following entries in the TOXDATA:
- 0.18%
- Low
- 27.7/57.4 ai
- 359 in air
- 359 in air
- 1083 in air
- 23 in vapor
- Exlclude the following entries with CHEMICAL
- Malathion Fyfanon 440g/L formulation
- Pyraclostrobin/Triticonazole BAS 673-01F mixture
- For those entries with TOXDATA including “form”, “product”, correct TOXDATA value by the “AI” column (which is the % active ingredient in the formulation)
unique(toxdata$DOSETYPE) beetox_OPP <- toxdata %>% filter(COMMONNAME=="Honey bee" & DOSETYPE %in% c("LC50", "LD50") & TGL=="") %>% filter(TOXLEVEL %in% c("UGB")) %>% filter(STUDYTIME %in% c("48 hr","24 hr","3 D", "44 hr","48 H","48 Hr") ) %>% filter(!(CHEMICAL %in% c("Malathion Fyfanon 440g/L formulation","Pyraclostrobin/Triticonazole BAS 673-01F mixture"))) dim(beetox_OPP) beetox_OPP <- beetox_OPP %>% filter(!(TOXICITY %in% c("0.18%", "Low", "27.7/57.4 ai", "359 in air", "1083 in air", "23 in vapor"))) id1 <- grep("/",beetox_OPP$AI) beetox_OPP[id1,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Mixtures and Formulations, mixtures will be excluded; Replace g/L assuming density = 1. So, 100g/L = 10%, 200g/L = 20%, etc.")%>%kableExtra::kable_paper() # [1] "95/5" "34/18" "56/6.2" "14/28" "200g/L" "100g/L" "20.6/20" "20.6/20" "1/7.5" "1/7.5" "13.3/5" #[12] "0.5/0.1" "58/6" "6/58" "17/26" "1.4/2.1" ### Remove all "/" entries since they are mixtures. ## Replace g/L assuming density = 1. So, 100g/L = 10%, 200g/L = 20%, etc. beetox_OPP <-beetox_OPP %>% mutate(AI=plyr::mapvalues(AI,from=c("200g/L","100g/L"),to=c("20","10"))) id1 <- grep("/",beetox_OPP$AI) beetox_OPP <- beetox_OPP[-id1,] grep("/",beetox_OPP$AI) id <- c(grep("form",beetox_OPP$TOXICITY),grep("product",beetox_OPP$TOXICITY)) grep("/",beetox_OPP$AI[id]) ### There is no form or product anymore!
id2 <- c(grep("WG",beetox_OPP$AI),grep("EC",beetox_OPP$AI),grep("SC",beetox_OPP$AI)) beetox_OPP[id2,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Formulations: Take the number as %")%>%kableExtra::kable_paper() ## For these ones remove the EC/SD or WG and use the nubmer as percentage? beetox_OPP <- beetox_OPP %>% mutate(ai=gsub("WG","",AI)) %>% mutate(ai=gsub("SC","",ai)) %>% mutate(ai=gsub("EC","",ai)) ## beetox_OPP[id2,"ai"] ## Now
Consider all the values that cannot be converted to a %.
id3 <-which(is.na(as.numeric(beetox_OPP$ai))) unique(beetox_OPP$AI[id3]) beetox_OPP[id3,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Other non-numeric AI entries")%>%kableExtra::kable_paper()
For Tech grade make the ai % 100%. Keep all "N.R" as 100%.
###### beetox_OPP$ai[grep("Tech",beetox_OPP$AI)] <- 100 beetox_OPP$ai[grep("tech",beetox_OPP$AI)] <- 100 beetox_OPP$ai[grep("N.R",beetox_OPP$AI)] <- 100 ################
id4 <- beetox_OPP$AI %in% (unique(beetox_OPP$AI[id3])[-c(1,4)]) beetox_OPP[id4,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="N.R, FORM, ISOM: NR being kept as 100%, other s excluded")%>%kableExtra::kable_paper()
Exclude all other.
id5 <- unique(c(grep("FORM",beetox_OPP$AI),grep("Form",beetox_OPP$AI),grep("ISOM",beetox_OPP$AI))) beetox_OPP <- beetox_OPP[-id5,] ### could be -id4
Check if ai is included in the Toxicity column
id6 <- unique(c(grep("ai",beetox_OPP$TOXICITY),grep("as",beetox_OPP$TOXICITY))) beetox_OPP[id6,c("CHEMICAL","AI","TOXICITY")] %>% kable(.,caption="Reported as AI or AS, but with AI entries not 100%, changed to 100%")%>%kableExtra::kable_paper() beetox_OPP$ai[id6] <- 100
beetox_OPP$ai <- as.numeric(beetox_OPP$ai) sum(is.na(beetox_OPP$ai))
## For those entries with TOXDATA including “form”, “product”, correct TOXDATA value by the “AI” column (which is the % active ingredient in the formulation) beetox_OPP <- beetox_OPP %>% mutate(tox=as.numeric(unlist(lapply(str_split(TOXICITY," "),function(x)x[1])))) summary(beetox_OPP$tox) #beetox_OPP$tox[beetox_OPP$TOXLEVEL=="PPB"] <- beetox_OPP$tox[beetox_OPP$TOXLEVEL=="PPB"]/1000 #beetox_OPP$TOXLEVEL[beetox_OPP$TOXLEVEL=="PPB"] <- "PPM" beetox_OPP$tox_ai <- beetox_OPP$tox*beetox_OPP$ai/100 sd(beetox_OPP$tox,na.rm = T) / mean(beetox_OPP$tox,na.rm = T) ## [1] 4.669772 beetox_OPP <- beetox_OPP %>% filter(!is.na(tox)) tmp <- beetox_OPP %>% group_by(CAS_NO)%>% summarise(ncas=n()) ## 159 CAS beetox_OPP <- left_join(beetox_OPP,tmp) beetox_OPP <- beetox_OPP%>% filter(ncas>3) length(unique(beetox_OPP$CAS_NO)) ## 45 CAS ==> 23 cas
library(skimr) cv <- function(x,na.rm=T){ sd(x,na.rm = na.rm)/mean(x,na.rm = na.rm) } my_skim <- skim_with(numeric = sfl(cv,nobs=length))
Not multiplied by AI column
cv_sum <- beetox_OPP %>% group_by(CAS_NO) %>% summarise(mean_LD=mean(tox),sd_LD=sd(tox),cv=sd_LD/mean_LD) cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="Summary stats for overall CAS LD mean, sd, and CV ")%>%kableExtra::kable_classic()
beetox_OPP %>% left_join(casinfo_OPP,by=c("CAS_NO"="INPUT"))%>% group_by(CAS_NO,PREFERRED_NAME) %>% my_skim(tox) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table: Summary stats for LD mean, sd, and CV for each CAS</strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
Multiplied by AI column
cv_sum <- beetox_OPP %>% group_by(CAS_NO) %>% summarise(mean_LD=mean(tox_ai),sd_LD=sd(tox_ai),cv=sd_LD/mean_LD) cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="Summary stats for overall CAS LD mean, sd, and CV ")%>%kableExtra::kable_classic()
beetox_OPP %>% left_join(casinfo_OPP,by=c("CAS_NO"="INPUT"))%>% group_by(CAS_NO,PREFERRED_NAME) %>% my_skim(tox_ai) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table: Summary stats for LD mean, sd, and CV for each CAS</strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
Ecotox DB Data
Assuming all ug/bee are ai ug/bee
Data.2$CAS.Number <- webchem::as.cas(Data.2$CAS.Number) #Data.3 <- left_join(Data.2,Data.SSD%>%mutate(CAS.Number=as.character(CAS.))) Data.3 <- left_join(Data.2,casinfo,by=c("CAS.Number"="INPUT")) cv_sum <- Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% summarise(mean_LD=mean(Observed.Response.Mean),sd_LD=sd(Observed.Response.Mean),cv=sd_LD/mean_LD) cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S1: Summary stats for overall CAS LD mean, sd, and CV </strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% my_skim(Observed.Response.Mean) %>% yank("numeric") %>% select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S2: Summary stats for LD mean, sd, and CV for each CAS</strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
Use the purity column and consider values tranformed to AI ug/bee data
Data.3 <- Data.3 %>%filter(!is.na(Data.3$Observed.Response.Mean.t)) cv_sum <- Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% summarise(mean_LD.t=mean(Observed.Response.Mean.t),sd_LD.t=sd(Observed.Response.Mean.t),cv=sd_LD.t/mean_LD.t) cv_sum %>% ungroup() %>% my_skim()%>% yank("numeric") %>% dplyr::select(-c( n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S1: Summary stats for overall CAS LD mean, sd, and CV </strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
Data.3 %>% group_by(CAS.Number,PREFERRED_NAME) %>% my_skim(Observed.Response.Mean.t) %>% yank("numeric") %>% dplyr::select(-c(skim_variable, n_missing,complete_rate)) %>% knitr::kable(.,digits=2,caption="<center><strong>Table S2: Summary stats for LD mean, sd, and CV for each CAS</strong></center>", escape = FALSE)%>%kableExtra::kable_classic()
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.