Coronavirus transmission risk increases along wildlife supply chains for human consumption in Viet Nam
Nguyen Quynh Huong, Nguyen Thi Thanh Nga, Nguyen Thu Thuy, Nguyen Van Long, Alice Latinne, Mathieu Pruvot, Nguyen Thanh Phuong, Le Tin Vinh Quang, Bach Duc Luu, Vo Van Hung, Nguyen Thi Lan, Nguyen Thi Hoa, Nguyen Tung, Van Dang Ky, Scott I. Roberton, Hoang Bich Thuy, Nguyen Van Long, Martin Gilbert, Leanne Wicker, Jonna A. K. Mazet, Christine Kreuder Johnson, Tracey Goldstein, Alex Tremeau-Bravard, Victoria Ontiveros, Damien O. Joly, Chris Walzer, Amanda E. Fine, and Sarah H. Olson
5/29/2020
Load Data
Load data from dryad repository and select one of the access options:
temp <- tempfile()
#download.file("https://doi.org/10.5061/dryad.7h44j0zrj",temp, mode="wb") #for use once the data set is public
unzip(temp, "VietnamWithInterpretations_20200323_dryad_v2.csv")
H1 <- read.csv("VietnamWithInterpretations_20200323_dryad_v2.csv")
#Alternately download .csv file here and move to your local directory - https://datadryad.org/stash/share/pk3wVUxFNzTuCYZ9t8haKRPmx7V8YhTDBuHpG8JJ9kU - for use during PLOS One review process
H1 <- read.csv("VietnamWithInterpretations_20200323_dryad_v2.csv")
head(H1)## Country SiteName District StateProv SiteLatitude
## 1 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## 2 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## 3 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## 4 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## 5 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## 6 Vietnam CPCP civet restaurant confiscation Nho Quan Ninh Binh 20.32
## SiteLongitude EventName DomesticAnimals
## 1 105.61 Cuc Phuong NP civet restaurant confiscation
## 2 105.61 Cuc Phuong NP civet restaurant confiscation
## 3 105.61 Cuc Phuong NP civet restaurant confiscation
## 4 105.61 Cuc Phuong NP civet restaurant confiscation
## 5 105.61 Cuc Phuong NP civet restaurant confiscation
## 6 105.61 Cuc Phuong NP civet restaurant confiscation
## AnimalID TaxaGroup SpeciesScientificName CommonName
## 1 VN11A0002 Carnivores Paradoxurus hermaphroditus Common Palm Civet
## 2 VN11A0002 Carnivores Paradoxurus hermaphroditus Common Palm Civet
## 3 VN11A0002 Carnivores Paradoxurus hermaphroditus Common Palm Civet
## 4 VN11A0002 Carnivores Paradoxurus hermaphroditus Common Palm Civet
## 5 VN11A0002 Carnivores Paradoxurus hermaphroditus Common Palm Civet
## 6 VN11A0001 Carnivores Paguma larvata Masked Palm Civet
## CommonNameFieldMorphology IDCertainty Order Family Genus
## 1 Common Palm Civet field ID certain CARNIVORA VIVERRIDAE Paradoxurus
## 2 Common Palm Civet field ID certain CARNIVORA VIVERRIDAE Paradoxurus
## 3 Common Palm Civet field ID certain CARNIVORA VIVERRIDAE Paradoxurus
## 4 Common Palm Civet field ID certain CARNIVORA VIVERRIDAE Paradoxurus
## 5 Common Palm Civet field ID certain CARNIVORA VIVERRIDAE Paradoxurus
## 6 Masked Palm Civet field ID certain CARNIVORA VIVERRIDAE Paguma
## SampleDate SpecimenID SpecimenType TestType TestRequested
## 1 4/22/11 HUA 52 red blood cells Conventional PCR Coronaviruses
## 2 4/22/11 HUA 55 oral swab Conventional PCR Coronaviruses
## 3 4/22/11 HUA 56 rectal swab Conventional PCR Coronaviruses
## 4 4/22/11 HUA 57 rectal swab Conventional PCR Coronaviruses
## 5 4/22/11 HUA 58 oral swab Conventional PCR Coronaviruses
## 6 4/22/11 HUA 51 red blood cells Conventional PCR Coronaviruses
## TestRequestedProtocol ConfirmationResult Sequence GenbankAccessionNumber
## 1 Quan | Watanabe Negative NULL NULL
## 2 Quan | Watanabe Negative NULL NULL
## 3 Quan | Watanabe Negative NULL NULL
## 4 Quan | Watanabe Negative NULL NULL
## 5 Quan | Watanabe Negative NULL NULL
## 6 Quan | Watanabe Negative NULL NULL
## Virus VirusGroup
## 1 NULL NULL
## 2 NULL NULL
## 3 NULL NULL
## 4 NULL NULL
## 5 NULL NULL
## 6 NULL NULLcolnames(H1)## [1] "Country" "SiteName"
## [3] "District" "StateProv"
## [5] "SiteLatitude" "SiteLongitude"
## [7] "EventName" "DomesticAnimals"
## [9] "AnimalID" "TaxaGroup"
## [11] "SpeciesScientificName" "CommonName"
## [13] "CommonNameFieldMorphology" "IDCertainty"
## [15] "Order" "Family"
## [17] "Genus" "SampleDate"
## [19] "SpecimenID" "SpecimenType"
## [21] "TestType" "TestRequested"
## [23] "TestRequestedProtocol" "ConfirmationResult"
## [25] "Sequence" "GenbankAccessionNumber"
## [27] "Virus" "VirusGroup"Clean up data/variables
Variable names
names(H1) <- to_snake_case(names(H1))Dates
H1$sample_date<-as.character(H1$sample_date)
H1$sample_date<-as.POSIXct(strptime(H1$sample_date, "%m/%d/%y", tz="Asia/bangkok"))
H1$sample_month<-month(H1$sample_date)Note: The sample dates that are earlier than 2011 correspond to archived samples that were obtained from Tam Dao Bear rescue center
Creating primary interface group
Establish primary interface groups based on site names.
H1$primary_interface_group<-NA
H1$primary_interface_group<-as.character(rep("captive facilities", n=nrow(H1)))
H1$primary_interface_group[is.element(H1$site_name, c("Rat market CT - Dong Thap","Rat market LVO - Dong Thap", "Rat market LVU - Dong Thap", "Rat market MX1 - Soc Trang", "Rat market MX2 - Soc Trang", "Rat market N5(1) - Soc Trang", "Rat market N5(2) - Soc Trang", "Rat market SD - Dong Thap"))] <-'trader'
H1$primary_interface_group[is.element(H1$site_name, c("Rat market CLC1 - Dong Thap", "Rat market CLC2 - Dong Thap", "Rat market CLC3 - Dong Thap", "Rat market CLD - Dong Thap", "Rat market HND - Dong Thap", "Rat market HNT - Dong Thap", "Rat market TB1 - Dong Thap", "Rat market TB2 - Dong Thap", "Rat market TH - Dong Thap", "Rat market TM1 - Dong Thap", "Rat market TM2 - Dong Thap", "Rat market TN - Dong Thap", "Rat market TT - Soc Trang", "Wet market (Whole sale sellers) - Soc Trang"))] <-'large market'
H1$primary_interface_group[is.element(H1$site_name, c("CPCP civet restaurant confiscation","Rat restaurant ST - Soc Trang", "Restaurant - Soc Trang", "Lam Dong FPD"))] <-'restaurant'
H1$primary_interface_group[is.element(H1$site_name, c("Bat pagoda - Soc Trang"))] <-'natural roost'
H1$primary_interface_group[is.element(H1$site_name, c("Bat roost and quano farm - Soc Trang", "Bat roost and quano farm CL - Dong Thap", "Bat roost and quano farm CLD2 - Soc Trang", "Bat roost and quano farm CLD3 - Soc Trang", "Bat roost and quano farm CLD5 - Soc Trang", "Bat roost and quano farm CLD6 - Soc Trang", "Bat roost and quano farm CLD7 - Soc Trang", "Bat roost and quano farm CT - Dong Thap", "Bat roost and quano farm HN - Dong Thap", "Bat roost and quano farm LP - Soc Trang", "Bat roost and quano farm LV - Dong Thap", "Bat roost and quano farm TM - Dong Thap", "Bat roost and quano farm CLD1 - Soc Trang", "Bat roost and quano farm CLD4 - Soc Trang", "Bat roost and quano farm MX - Soc Trang", "Bat roost and quano farm N5(1) - Soc Trang", "Bat roost and quano farm N5(2) - Soc Trang"))] <-'bat guano farm'
H1$primary_interface_group[is.element(H1$site_name, c("Farm 1 - Bien Hoa - DN", "Farm 10 - Vinh Cuu - DN", "Farm 12 - Nhon Trach - DN", "Farm 13 - Nhon Trach - DN", "Farm 14 - Long Thanh - DN", "Farm 15 - Long Thanh - DN", "Farm 16 - Long Thanh - DN", "Farm 17 - Long Thanh - DN", "Farm 18 - Cam My - DN", "Farm 19 - Cam My - DN", "Farm 2 - Bien Hoa - DN", "Farm 20 - Cam My - DN", "Farm 21 - Xuan Loc - DN", "Farm 22 - Xuan Loc - DN", "Farm 23 - Xuan Loc - DN", "Farm 24 - Xuan Loc - DN", "Farm 25 - Long Khanh - DN", "Farm 26 - Xuan Loc - DN", "Farm 28 - Dinh Quan - DN", "Farm 29 - Tan Phu - DN", "Farm 3 - Bien Hoa - DN", "Farm 4 - Thong Nhat - DN", "Farm 5 - Thong Nhat - DN", "Farm 6 - Trang Bom - DN", "Farm 8 - Vinh Cuu - DN", "Farm 9 - Vinh Cuu - DN", "Wildlife farm 1 - Dong Nai", "Wildlife farm 3 - Dong Nai", "Wildlife farm 4 - Dong Nai"))] <-'wildlife farm'
H1$primary_interface_group <- as.factor(H1$primary_interface_group)
H1$primary_interface_group <- factor(H1$primary_interface_group, levels = c("trader", "large market", "restaurant", "natural roost","bat guano farm","wildlife farm","captive facilities"))Cleaning genus
Add family for some missing genus in the data (couple squirrels and unidentified viverridae). Also added the captive facility for the bears and macaques, even though they tested negative (avoid missing values)
H1$new_genus<-NA
H1$new_genus<-as.character(H1$genus)
H1<-H1 %>%
mutate(new_genus=replace(new_genus,H1$genus=="NULL" & species_scientific_name=="Viverridae","Viverridae"))
H1<-H1 %>%
mutate(new_genus=replace(new_genus,H1$genus=="NULL" & species_scientific_name=="Sciuridae","Sciuridae"))
H1<-H1 %>%
mutate(new_genus=replace(new_genus,H1$genus=="NULL" & species_scientific_name=="Chiroptera","Microchiroptera"))
H1<-H1 %>%
mutate(new_genus=replace(new_genus,primary_interface_group=="trader" & family=="MURIDAE","Field Rats"))
H1<-H1 %>%
mutate(new_genus=replace(new_genus,primary_interface_group=="large market" & family=="MURIDAE","Field Rats"))
H1<-H1 %>%
mutate(new_genus=replace(new_genus,primary_interface_group=="restaurant" & family=="MURIDAE","Field Rats"))Create variable that lists viruses associated with each animal id to determine coinfection status
Each row in the data set is a test and virus sequence result so need to add a field that summarizes all viruses found in a single individual to establish co-infection status
tempH1<-as.data.frame(H1)
tempH1$virus<-as.character(tempH1$virus)
tempH1$virus[tempH1$virus == "NULL"] <- NA #replace NULL with NA
temp1 <- tempH1 %>%
drop_na(virus) %>%
group_by(animal_id) %>%
summarize(virus_ind = (paste0(unique(virus_group), collapse=" | "))) %>%
as.data.frame()
temp2 <- tempH1 %>%
group_by(animal_id, specimen_id) %>%
summarize(virus_specimen = paste(unique(virus_group), collapse=" | "))%>% as.data.frame()
temp3<-full_join(temp1, temp2, by="animal_id")
temp4<-subset(temp3, select=-animal_id)
H1<-full_join(temp4, H1, by="specimen_id")Define wet and dry seasons based on sampling date
The annual wet season in southern Vietnam occurs from May 1st through November 30th, and the dry season from December 1st through April 30th.
H1$hseason <- NA
H1$hseason <- ifelse(H1$sample_month %in% c(12,1:4), "Dry","Wet")
table(H1$sample_month,H1$hseason) ##
## Dry Wet
## 1 643 0
## 3 850 0
## 4 82 0
## 6 0 100
## 8 0 46
## 9 0 9
## 10 0 2793
## 11 0 96
## 12 3 0Select just rodents and bats for futher analysis
H1<-subset(H1, is.element(taxa_group,c("Bats","Rodents & Shrews")))
H1<-droplevels(H1)Materials and methods
Sampling locations
Sampling site description by interface type and province:
H1 %>%
group_by(primary_interface_group) %>%
summarize(site_counts=length(unique(site_name)))## # A tibble: 6 x 2
## primary_interface_group site_counts
## <fct> <int>
## 1 trader 8
## 2 large market 14
## 3 restaurant 2
## 4 natural roost 1
## 5 bat guano farm 17
## 6 wildlife farm 28H1 %>%
group_by(primary_interface_group, state_prov) %>%
summarize(site_counts=length(unique(site_name)))## # A tibble: 9 x 3
## # Groups: primary_interface_group [6]
## primary_interface_group state_prov site_counts
## <fct> <fct> <int>
## 1 trader Dong Thap 4
## 2 trader Soc Trang 4
## 3 large market Dong Thap 12
## 4 large market Soc Trang 2
## 5 restaurant Soc Trang 2
## 6 natural roost Soc Trang 1
## 7 bat guano farm Dong Thap 5
## 8 bat guano farm Soc Trang 12
## 9 wildlife farm Dong Nai 28Timing of sampling:
H1 %>%
group_by(hseason) %>%
summarize(site_count=length(unique(site_name)))## # A tibble: 2 x 2
## hseason site_count
## <chr> <int>
## 1 Dry 30
## 2 Wet 41min(H1$sample_date)## [1] "2013-01-11 +07"max(H1$sample_date)## [1] "2014-03-29 +07"Number of visits per site (and per season):
H1 %>%
group_by(site_name) %>%
summarize(site_visits=length(unique(event_name)))## # A tibble: 70 x 2
## site_name site_visits
## <fct> <int>
## 1 Bat pagoda - Soc Trang 3
## 2 Bat roost and quano farm - Soc Trang 1
## 3 Bat roost and quano farm CL - Dong Thap 1
## 4 Bat roost and quano farm CLD1 - Soc Trang 1
## 5 Bat roost and quano farm CLD2 - Soc Trang 1
## 6 Bat roost and quano farm CLD3 - Soc Trang 1
## 7 Bat roost and quano farm CLD4 - Soc Trang 1
## 8 Bat roost and quano farm CLD5 - Soc Trang 1
## 9 Bat roost and quano farm CLD6 - Soc Trang 1
## 10 Bat roost and quano farm CLD7 - Soc Trang 1
## # … with 60 more rowsH1 %>%
group_by(site_name,hseason) %>%
summarize(vists=length(unique(sample_date)))## # A tibble: 71 x 3
## # Groups: site_name [70]
## site_name hseason vists
## <fct> <chr> <int>
## 1 Bat pagoda - Soc Trang Dry 5
## 2 Bat pagoda - Soc Trang Wet 1
## 3 Bat roost and quano farm - Soc Trang Dry 1
## 4 Bat roost and quano farm CL - Dong Thap Wet 1
## 5 Bat roost and quano farm CLD1 - Soc Trang Wet 1
## 6 Bat roost and quano farm CLD2 - Soc Trang Wet 1
## 7 Bat roost and quano farm CLD3 - Soc Trang Wet 1
## 8 Bat roost and quano farm CLD4 - Soc Trang Wet 1
## 9 Bat roost and quano farm CLD5 - Soc Trang Wet 1
## 10 Bat roost and quano farm CLD6 - Soc Trang Wet 1
## # … with 61 more rowsIn most cases, there was only one visit per site, except for the bat pagoda. This should call for caution when looking at seasonal effect, as site and season are confounded (so season effect may just be a result of the timing of sampling across the different sites). Only the Soc Trang bat pagoda had sampling in two seasons and is suitable for assessment of seasonal effect.
Animal Sampling
Summary tables clarifying what sampling protocol was used in each interface/site.
H1 %>%
select(primary_interface_group,specimen_type) %>%
group_by(primary_interface_group) %>%
count(specimen_type) %>%
pivot_wider(names_from = specimen_type,values_from = n)## # A tibble: 6 x 12
## # Groups: primary_interface_group [6]
## primary_interfa… brain lung `oral swab` `small intestin… feces kidney
## <fct> <int> <int> <int> <int> <int> <int>
## 1 trader 50 94 308 232 NA NA
## 2 large market 90 102 589 443 26 122
## 3 restaurant NA 170 239 191 4 NA
## 4 natural roost NA NA 30 NA 90 NA
## 5 bat guano farm NA NA NA NA 624 NA
## 6 wildlife farm NA NA NA NA 796 NA
## # … with 5 more variables: `rectal swab` <int>, `urine/urogenital swab` <int>,
## # spleen <int>, urine <int>, `environmental sample` <int>H1 %>%
select(site_name,specimen_type) %>%
group_by(site_name) %>%
count(specimen_type) %>%
pivot_wider(names_from = specimen_type,values_from = n)## # A tibble: 70 x 12
## # Groups: site_name [70]
## site_name feces `oral swab` `rectal swab` urine `environmental …
## <fct> <int> <int> <int> <int> <int>
## 1 Bat pago… 90 30 36 4 NA
## 2 Bat roos… 16 NA NA NA NA
## 3 Bat roos… 19 NA NA NA NA
## 4 Bat roos… 12 NA NA NA NA
## 5 Bat roos… 20 NA NA NA NA
## 6 Bat roos… 14 NA NA NA NA
## 7 Bat roos… 10 NA NA NA NA
## 8 Bat roos… 24 NA NA NA NA
## 9 Bat roos… 6 NA NA NA NA
## 10 Bat roos… 10 NA NA NA NA
## # … with 60 more rows, and 6 more variables: `urine/urogenital swab` <int>,
## # kidney <int>, `small intestine` <int>, brain <int>, lung <int>,
## # spleen <int>Results
Detection of coronavirus by animal taxa and interface
Paragraph 1 overview of specimens and animals
Number of samples:
length(unique(H1$specimen_id))## [1] 2164Number of animals:
length(unique(H1$animal_id)) #ID assigned in the field## [1] 1506Break-down of animal number and proportion of positive by taxonomic group:
H1 %>%
group_by(taxa_group) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100)## # A tibble: 2 x 4
## taxa_group ind_count ind_count_pos percent_pos
## <fct> <int> <int> <dbl>
## 1 Bats 375 238 63.5
## 2 Rodents & Shrews 1131 266 23.5Distribution of samples and animals by province
H1 %>%
group_by(state_prov) %>%
summarize(ind_count=length(unique(animal_id)), site_count=length(unique(site_name)))## # A tibble: 3 x 3
## state_prov ind_count site_count
## <fct> <int> <int>
## 1 Dong Nai 429 28
## 2 Dong Thap 565 21
## 3 Soc Trang 512 21S1 Table
Summary of all testing results by genus, interface, sub-interface, sample types, number specimens tested, number specimens tested positives, %positive and viral species.
S1Table <- H1 %>%
group_by(new_genus, primary_interface_group, specimen_type) %>%
summarize(site_count = length(unique(site_name)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
ind_count=length(unique(animal_id)),
percent_pos = ind_count_pos/ind_count*100,
IndivLevel= paste0(round(percent_pos,digits = 1), "% (",ind_count_pos,"/",ind_count,")"),
viruses = paste(unique(virus_specimen), collapse=", "))
S1Table## # A tibble: 32 x 9
## # Groups: new_genus, primary_interface_group [10]
## new_genus primary_interfa… specimen_type site_count ind_count_pos ind_count
## <chr> <fct> <fct> <int> <int> <int>
## 1 Cynopter… natural roost oral swab 1 0 1
## 2 Cynopter… natural roost rectal swab 1 0 2
## 3 Field Ra… trader brain 1 1 25
## 4 Field Ra… trader lung 4 6 47
## 5 Field Ra… trader oral swab 8 28 154
## 6 Field Ra… trader small intest… 7 14 116
## 7 Field Ra… large market brain 5 4 45
## 8 Field Ra… large market feces 1 0 13
## 9 Field Ra… large market kidney 8 3 61
## 10 Field Ra… large market lung 3 4 51
## # … with 22 more rows, and 3 more variables: percent_pos <dbl>,
## # IndivLevel <chr>, viruses <chr>#write.csv(S1Table,'S1Table.csv')Figure 4
Build out map side of figure
Get Vietnam map here from geoBoundaries v. 3.0.0 (tohttps://www.geoboundaries.org; Runfola et al. 2020) and https://vietnam.opendevelopmentmekong.net/terms-of-use/ (ODM)
temp <- tempfile()
download.file("https://geoboundaries.org//data//geoBoundariesSSCGS-3_0_0//VNM//ADM1//geoBoundariesSSCGS-3_0_0-VNM-ADM1-all.zip",temp, mode="wb")
unzip(temp, "geoBoundariesSSCGS-3_0_0-VNM-ADM1-shp.zip")
unzip("geoBoundariesSSCGS-3_0_0-VNM-ADM1-shp.zip", exdir='.')
vprovinces <- st_read("geoboundariesSSCGS-3_0_0-VNM-ADM1.shp")## Reading layer `geoboundariesSSCGS-3_0_0-VNM-ADM1' from data source `/Users/olson/R_Githubs/huong_paper/geoboundariesSSCGS-3_0_0-VNM-ADM1.shp' using driver `ESRI Shapefile'
## Simple feature collection with 63 features and 5 fields
## geometry type: MULTIPOLYGON
## dimension: XY
## bbox: xmin: 102.1439 ymin: 8.564999 xmax: 109.4593 ymax: 23.39253
## CRS: 4326unlink(temp)
newdataProjected <- as(vprovinces, "Spatial")Add in color coding for each province
newdataProjected@data$colorchoice <- rep("white", times=length(newdataProjected@data$shapeName))
newdataProjected@data$colorchoice[newdataProjected@data$shapeName=="Dong Nai"] <-'grey'
newdataProjected@data$colorchoice[newdataProjected@data$shapeName=="Dong Thap"] <-'grey'
newdataProjected@data$colorchoice[newdataProjected@data$shapeName=="Soc Trang"] <-'grey'Convert to sf object
newdataProjected_sf <- st_as_sf(newdataProjected)Create data for labeling the 3 provinces
newlabeldata <- st_centroid(newdataProjected_sf[newdataProjected_sf$shapeName %in% c("Dong Nai","Dong Thap", "Soc Trang"),])create map
newggViet <-ggplot(data=newdataProjected_sf)+
geom_sf(aes(fill=colorchoice), color='black',size =0.5) +
scale_fill_identity("Province status",
labels=c("Sampling","No sampling"),
guide="legend")+
geom_text_repel(data = newlabeldata,
aes(st_coordinates(newlabeldata)[,1],st_coordinates(newlabeldata)[,2],label = shapeName),
#fontface = "bold",
nudge_x = c(-2, -2, 1),
nudge_y = c(2, 1, -2),
size = 10*0.352777778) +
coord_sf() +
theme(legend.text = element_text(size=10),
legend.position = c(0.23,0.5),
rect = element_blank(),
axis.ticks = element_blank(),
plot.margin=unit(c(0,0,0,0.5),"cm"),
axis.text.y = element_blank(),
axis.text.x = element_blank(),
axis.title.y=element_blank(),
axis.title.x=element_blank())
print(newggViet)
Build out panel side of figure
Create data and panel for each province. First, clarify taxomic group for the figure:
H1$plot_taxa <-as.character(H1$species_scientific_name)
H1$plot_taxa[H1$species_scientific_name=="Chiroptera"] <-'Microchiroptera'
H1$plot_taxa[H1$primary_interface_group=="trader"] <-'Field rats'
H1$plot_taxa[H1$primary_interface_group=="large market"] <-'Field rats'
H1$plot_taxa[H1$primary_interface_group=="restaurant"] <-'Field rats'
H1$plot_taxa[H1$genus=="Pteropus"] <-"Pteropodidae"
H1$plot_taxa[H1$genus=="Cynopterus"] <-"Pteropodidae"
H1$plot_taxa[H1$genus=="Rhizomys"] <-"Rhizomys"
H1$plot_taxa <- as.factor(H1$plot_taxa)
H1$plot_taxa <- factor(H1$plot_taxa, levels = c("Field rats", "Hystrix brachyura", "Rhizomys", "Rattus argentiventer", "Sciuridae", "Microchiroptera", "Pteropodidae"))
H1$primary_interface_group_2 <- H1$primary_interface_group
levels(H1$primary_interface_group_2) <- c("Rat trader", "Large\nmarket", "Restaurant", "Natural\nbat\nroost", "Bat\nguano\nfarm","Wildlife\nfarm")
H1$primary_interface_group_2 <- factor(H1$primary_interface_group_2, levels=c("Rat trader", "Large\nmarket", "Restaurant", "Wildlife\nfarm", "Natural\nbat\nroost", "Bat\nguano\nfarm"))create summary data for figure
F4tab<- H1 %>%
group_by(state_prov, primary_interface_group_2, plot_taxa) %>%
summarize(ind_count=length(unique(animal_id)))Bar chart panels
Fig4<-ggplot(F4tab, aes(x=primary_interface_group_2, y=ind_count, fill=plot_taxa)) +
scale_fill_brewer(palette = "Set1") +
geom_bar(stat='identity') +
facet_wrap(~state_prov, ncol=1) +
ylab("Individual count") +
xlab("Sub-interface or interface") +
labs(fill = "Taxa group") +
theme_bw() +
theme(axis.text=element_text(size=8),
# axis.text.x = element_text(angle=45, hjust=1),
axis.title = element_text(face="bold"),
legend.text =element_text(size=8),
legend.title =element_text(size=10),
strip.text = element_text(size = 8),
plot.margin=unit(c(0.25,0,0.25,0),"cm"))
print(Fig4)
combine map and bar charts for final Fig 4:
grid.arrange(newggViet, Fig4, ncol = 3,
layout_matrix = cbind(c(1,1), c(2,2), c(2,2)))
#tiff("Fig4.tif", height = 6, width = 7.4, units = 'in', res=300)
#grid.arrange(newggViet, Fig4, ncol = 3, layout_matrix = cbind(c(1,1), c(2,2), c(2,2)))
#dev.off()Paragraph 2: Number of sites and proportion of positive sites
Number of sites
length(unique(H1$site_name))## [1] 70Number of sites where coronavirus was detected
length(unique(H1[H1$confirmation_result=="Positive","site_name"]))## [1] 58Compile number of sites with positive samples and site-level prevalence by interface and taxa_group
H2.1i<- H1%>%
group_by(taxa_group,primary_interface_group,site_name) %>%
summarize(pos_site=sum(sum(confirmation_result=="Positive")>0))
H2.1i %>%
group_by(taxa_group,primary_interface_group) %>%
summarize(site_count=length(unique(site_name)),pos_sites=sum(pos_site),siteprev=pos_sites/site_count)## # A tibble: 6 x 5
## # Groups: taxa_group [2]
## taxa_group primary_interface_group site_count pos_sites siteprev
## <fct> <fct> <int> <int> <dbl>
## 1 Bats natural roost 1 1 1
## 2 Bats bat guano farm 17 16 0.941
## 3 Rodents & Shrews trader 8 8 1
## 4 Rodents & Shrews large market 14 14 1
## 5 Rodents & Shrews restaurant 2 2 1
## 6 Rodents & Shrews wildlife farm 28 17 0.607Table 1
This code is producing different sections of Table 1
Main structure:
T1<- H1 %>%
group_by(taxa_group, primary_interface_group, new_genus) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100,
site_count=length(unique(site_name)),
pos_sites=length(unique(site_name[confirmation_result=="Positive"])),
siteprev=pos_sites/site_count*100,
SiteLevel= paste0(round(siteprev,digits = 1), "% (",pos_sites,"/",site_count,")"),
IndivLevel= paste0(round(percent_pos,digits = 1), "% (",ind_count_pos,"/",ind_count,")"),
viruses = paste0(unique(virus_ind), collapse=", "))%>%
select(-c(4:9)) %>%
arrange(desc(taxa_group))
T1## # A tibble: 10 x 6
## # Groups: taxa_group, primary_interface_group [6]
## taxa_group primary_interfac… new_genus SiteLevel IndivLevel viruses
## <fct> <fct> <chr> <chr> <chr> <chr>
## 1 Rodents & … trader Field Rats 100% (8/8) 20.7% (39… Murine corona…
## 2 Rodents & … large market Field Rats 100% (14/… 32% (116/… Murine corona…
## 3 Rodents & … restaurant Field Rats 100% (2/2) 55.6% (84… Murine corona…
## 4 Rodents & … wildlife farm Hystrix 47.8% (11… 6% (20/33… Bat coronavir…
## 5 Rodents & … wildlife farm Rattus 100% (1/1) 100% (1/1) Bat coronavir…
## 6 Rodents & … wildlife farm Rhizomys 45.5% (5/… 6.2% (6/9… Infectious br…
## 7 Rodents & … wildlife farm Sciuridae 0% (0/1) 0% (0/1) NA
## 8 Bats natural roost Cynopterus 0% (0/1) 0% (0/2) NA
## 9 Bats natural roost Pteropus 100% (1/1) 6.7% (4/6… PREDICT_CoV-1…
## 10 Bats bat guano farm Microchir… 94.1% (16… 74.8% (23… Bat coronavir…Number of individuals found with each virus:
T1v<- H1 %>%
group_by(taxa_group, primary_interface_group, new_genus, virus_group) %>%
summarize(ind_count_virus=length(unique(animal_id))) %>%
mutate(virus_ind_count=paste0(virus_group, " (", ind_count_virus,")"))%>%
group_by(taxa_group, primary_interface_group, new_genus) %>%
summarize(viral_species=paste0(virus_ind_count ,collapse = ', ')) %>%
arrange(desc(taxa_group))
T1v## # A tibble: 10 x 4
## # Groups: taxa_group, primary_interface_group [6]
## taxa_group primary_interface_… new_genus viral_species
## <fct> <fct> <chr> <chr>
## 1 Rodents & S… trader Field Rats Longquan Aa mouse coronavirus (5…
## 2 Rodents & S… large market Field Rats Longquan Aa mouse coronavirus (3…
## 3 Rodents & S… restaurant Field Rats Longquan Aa mouse coronavirus (2…
## 4 Rodents & S… wildlife farm Hystrix Bat coronavirus 512/2005 (19), I…
## 5 Rodents & S… wildlife farm Rattus Bat coronavirus 512/2005 (1), NU…
## 6 Rodents & S… wildlife farm Rhizomys Bat coronavirus 512/2005 (5), In…
## 7 Rodents & S… wildlife farm Sciuridae NULL (1)
## 8 Bats natural roost Cynopterus NULL (2)
## 9 Bats natural roost Pteropus NULL (59), PREDICT_CoV-17 (3), P…
## 10 Bats bat guano farm Microchir… Bat coronavirus 512/2005 (216), …Coinfections column
coinfects<-H1[grep('\\|',H1$virus_ind),]
T1coinf<-coinfects %>%
group_by(taxa_group, primary_interface_group, new_genus) %>%
summarize(ind_count=length(unique(animal_id))) %>%
arrange(desc(taxa_group))
T1coinf## # A tibble: 4 x 4
## # Groups: taxa_group, primary_interface_group [4]
## taxa_group primary_interface_group new_genus ind_count
## <fct> <fct> <chr> <int>
## 1 Rodents & Shrews trader Field Rats 2
## 2 Rodents & Shrews large market Field Rats 18
## 3 Rodents & Shrews restaurant Field Rats 6
## 4 Bats bat guano farm Microchiroptera 21Paragraph 3 Field rodents
Subset the data for rodents:
H4<-subset(H1, taxa_group=="Rodents & Shrews")
length(unique(H4$animal_id))## [1] 1131Proportion positive at site and animal level by province:
H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
group_by(state_prov) %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [2 x 8]
## Groups: <by row>
##
## # A tibble: 2 x 8
## state_prov site_count pos_sites siteprev ind_count ind_count_pos percent_pos
## <fct> <int> <int> <dbl> <int> <int> <dbl>
## 1 Dong Thap 16 16 100 373 129 34.6
## 2 Soc Trang 8 8 100 329 110 33.4
## # … with 1 more variable: confint <chr>Overall proportion of positive field rats:
H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [1 x 7]
## Groups: <by row>
##
## # A tibble: 1 x 7
## site_count pos_sites siteprev ind_count ind_count_pos percent_pos confint
## <int> <int> <dbl> <int> <int> <dbl> <chr>
## 1 24 24 100 702 239 34.0 30.6 - 37.7proportion of positive field rats by site:
H4bysite <- H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
group_by(site_name) %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))
H4bysite## Source: local data frame [24 x 8]
## Groups: <by row>
##
## # A tibble: 24 x 8
## site_name site_count pos_sites siteprev ind_count ind_count_pos percent_pos
## <fct> <int> <int> <dbl> <int> <int> <dbl>
## 1 Rat mark… 1 1 100 9 5 55.6
## 2 Rat mark… 1 1 100 20 9 45
## 3 Rat mark… 1 1 100 15 5 33.3
## 4 Rat mark… 1 1 100 20 5 25
## 5 Rat mark… 1 1 100 26 6 23.1
## 6 Rat mark… 1 1 100 26 2 7.69
## 7 Rat mark… 1 1 100 20 7 35
## 8 Rat mark… 1 1 100 22 11 50
## 9 Rat mark… 1 1 100 18 4 22.2
## 10 Rat mark… 1 1 100 19 6 31.6
## # … with 14 more rows, and 1 more variable: confint <chr>H4bysite%>%
summarize(max_site=max(percent_pos),min_site=min(percent_pos))## # A tibble: 24 x 2
## max_site min_site
## <dbl> <dbl>
## 1 55.6 55.6
## 2 45 45
## 3 33.3 33.3
## 4 25 25
## 5 23.1 23.1
## 6 7.69 7.69
## 7 35 35
## 8 50 50
## 9 22.2 22.2
## 10 31.6 31.6
## # … with 14 more rowsProportion of positive rodents by sub-interface type
H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
group_by(primary_interface_group) %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100)%>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [3 x 8]
## Groups: <by row>
##
## # A tibble: 3 x 8
## primary_interfa… site_count pos_sites siteprev ind_count ind_count_pos
## <fct> <int> <int> <dbl> <int> <int>
## 1 trader 8 8 100 188 39
## 2 large market 14 14 100 363 116
## 3 restaurant 2 2 100 151 84
## # … with 2 more variables: percent_pos <dbl>, confint <chr>Proportion of positive rodents by season
H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
group_by(hseason) %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),
siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100)%>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [2 x 8]
## Groups: <by row>
##
## # A tibble: 2 x 8
## hseason site_count pos_sites siteprev ind_count ind_count_pos percent_pos
## <chr> <int> <int> <dbl> <int> <int> <dbl>
## 1 Dry 2 2 100 130 29 22.3
## 2 Wet 22 22 100 572 210 36.7
## # … with 1 more variable: confint <chr>Testing effect of season and sub-interface on risk of infection of rodent along supply chain
Setting up the data for uni- and multi-variable models
H1st<-H1 %>%
filter(state_prov %in% c("Dong Thap", "Soc Trang") & primary_interface_group %in% c("trader","large market","restaurant")) %>%
group_by(new_genus,animal_id) %>%
summarize(coronavirus=length(unique(animal_id[confirmation_result=="Positive"])),
season = unique(hseason),
site_name = unique(site_name),
interface= unique(primary_interface_group))
H1st$season <- as.factor(H1st$season)
H1st$season <- factor(H1st$season, levels = c("Dry", "Wet"))
H1st$interface<- as.factor(H1st$interface)
H1st$interface<- factor(H1st$interface, levels = c("trader","large market","restaurant"))
H1st## # A tibble: 702 x 6
## # Groups: new_genus [1]
## new_genus animal_id coronavirus season site_name interface
## <chr> <fct> <int> <fct> <fct> <fct>
## 1 Field Rats VN13M0001 0 Dry Restaurant - Soc Trang restaurant
## 2 Field Rats VN13M0002 1 Dry Restaurant - Soc Trang restaurant
## 3 Field Rats VN13M0003 1 Dry Restaurant - Soc Trang restaurant
## 4 Field Rats VN13M0004 0 Dry Wet market (Whole sale se… large mar…
## 5 Field Rats VN13M0005 1 Dry Restaurant - Soc Trang restaurant
## 6 Field Rats VN13M0006 0 Dry Restaurant - Soc Trang restaurant
## 7 Field Rats VN13M0007 1 Dry Restaurant - Soc Trang restaurant
## 8 Field Rats VN13M0008 1 Dry Restaurant - Soc Trang restaurant
## 9 Field Rats VN13M0009 1 Dry Restaurant - Soc Trang restaurant
## 10 Field Rats VN13M0010 0 Dry Wet market (Whole sale se… large mar…
## # … with 692 more rowsUnivariate models for season and interface:
m_st.1<-glm(coronavirus~season, family = binomial, data = H1st)
summary(m_st.1)##
## Call:
## glm(formula = coronavirus ~ season, family = binomial, data = H1st)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9566 -0.9566 -0.9566 1.4157 1.7322
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.2478 0.2107 -5.923 3.16e-09 ***
## seasonWet 0.7033 0.2278 3.087 0.00202 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 900.44 on 701 degrees of freedom
## Residual deviance: 890.08 on 700 degrees of freedom
## AIC: 894.08
##
## Number of Fisher Scoring iterations: 4exp(cbind(OR = coef(m_st.1), confint(m_st.1)))## Waiting for profiling to be done...## OR 2.5 % 97.5 %
## (Intercept) 0.2871287 0.1867155 0.4277976
## seasonWet 2.0203848 1.3082077 3.2047957m_st.2<-glm(coronavirus~interface, family = binomial, data = H1st)
summary(m_st.2)##
## Call:
## glm(formula = coronavirus ~ interface, family = binomial, data = H1st)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.2748 -0.8775 -0.6819 1.0830 1.7736
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.3404 0.1799 -7.452 9.19e-14 ***
## interfacelarge market 0.5846 0.2122 2.755 0.00587 **
## interfacerestaurant 1.5665 0.2433 6.439 1.20e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 900.44 on 701 degrees of freedom
## Residual deviance: 854.25 on 699 degrees of freedom
## AIC: 860.25
##
## Number of Fisher Scoring iterations: 4exp(cbind(OR = coef(m_st.2), confint(m_st.2)))## Waiting for profiling to be done...## OR 2.5 % 97.5 %
## (Intercept) 0.261745 0.181572 0.3682482
## interfacelarge market 1.794249 1.192172 2.7436193
## interfacerestaurant 4.789897 2.993666 7.7802354Multivariable model:
m_st.4<-glmer(coronavirus~interface+season +(1|site_name), family = binomial, data = H1st)
summary(m_st.4)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: coronavirus ~ interface + season + (1 | site_name)
## Data: H1st
##
## AIC BIC logLik deviance df.resid
## 801.6 824.4 -395.8 791.6 697
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.4641 -0.6956 -0.3648 0.7019 3.1641
##
## Random effects:
## Groups Name Variance Std.Dev.
## site_name (Intercept) 0.4105 0.6407
## Number of obs: 702, groups: site_name, 24
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.9838 0.6842 -4.361 1.29e-05 ***
## interfacelarge market 0.7892 0.3693 2.137 0.032567 *
## interfacerestaurant 2.3052 0.6441 3.579 0.000345 ***
## seasonWet 1.5927 0.6165 2.583 0.009788 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) intrfm intrfc
## intrfclrgmr -0.466
## intrfcrstrn -0.622 0.430
## seasonWet -0.902 0.134 0.469check if any site was sampled more than once
H1 %>%
filter(state_prov %in% c("Dong Thap", "Soc Trang") & primary_interface_group %in% c("trader","large market","restaurant")) %>%
group_by(site_name) %>%
summarize(n_visit=length(unique(event_name)))## # A tibble: 24 x 2
## site_name n_visit
## <fct> <int>
## 1 Rat market CLC1 - Dong Thap 1
## 2 Rat market CLC2 - Dong Thap 1
## 3 Rat market CLC3 - Dong Thap 1
## 4 Rat market CLD - Dong Thap 1
## 5 Rat market CT - Dong Thap 1
## 6 Rat market HND - Dong Thap 1
## 7 Rat market HNT - Dong Thap 1
## 8 Rat market LVO - Dong Thap 1
## 9 Rat market LVU - Dong Thap 1
## 10 Rat market MX1 - Soc Trang 1
## # … with 14 more rowsSite only sampled once, so practically, both season and interface are site-level variables, and confounding effects with other site-level characteristics cannot be excluded. Although the GLMM is the right modeling approach for this data structure, caution required in the interpretation due to sampling design limitations
Table S2 Field rodent multi-variable model with site as a random effect
TS2 <- as.data.frame(exp(confint(m_st.4)))## Computing profile confidence intervals ...TS2$OR <- NA
TS2$OR[3:5] <- exp(summary(m_st.4)$coefficients[2:4,1])
TS2[3:5,c(3,1,2)]## OR 2.5 % 97.5 %
## interfacelarge market 2.201718 1.045097 4.751261
## interfacerestaurant 10.025683 2.665258 39.453799
## seasonWet 4.916909 1.375610 18.024858Figure 5 Graph representation of these results
H1$primary_interface_group_3 <- H1$primary_interface_group
levels(H1$primary_interface_group_3) <- c("Rat trader", "Large market", "Restaurant", "Natural bat roost", "Bat guano farm","Wildlife farm")
RodT_int <- H1 %>%
filter(state_prov %in% c("Dong Thap", "Soc Trang", "Dong Nai") & taxa_group=="Rodents & Shrews" & primary_interface_group_3 %in% c("Rat trader", "Large market", "Restaurant")) %>%
group_by(primary_interface_group_3) %>%
summarize(N=length(unique(animal_id)),
Positive=length(unique(animal_id[confirmation_result=="Positive"])),
Proportion = 100*Positive/N,
lab = paste0(round(Proportion,1), "%\n n=", N),
L_CI = round(100*prop.test(Positive,N, conf.level=0.95)$conf.int[1],1),
U_CI = round(100*prop.test(Positive,N, conf.level=0.95)$conf.int[2],1)) %>%
as.data.frame()
Fig5 <- ggplot(RodT_int,aes(primary_interface_group_3, Proportion)) +
geom_col(fill="lightgrey") +
geom_errorbar(aes(ymin = ifelse(L_CI<0,0,L_CI), ymax = U_CI), width = .5, color="grey", alpha=0.9) +
xlab("Field rat sub-interface") +
geom_text(aes(label = lab, hjust = .5, vjust= 0.50), position = position_dodge(width = 0),
angle = 0)+
theme_bw()
Fig5
#tiff("Fig5.tif", height = 4, width = 4, units = 'in', res=300)
#grid.arrange(Fig5, ncol = 1, layout_matrix = cbind(c(1,1)))
#dev.off()Paragraph 4 - Probability of detection with different sample types among positive field rats with more than one sample
Select animals id from the rodent trade that:
- have multiple sample types
- are positive
Rp<-H4 %>%
filter(confirmation_result=="Positive")
Rmsp<-H4 %>%
filter(primary_interface_group!="wildlife farm") %>%
group_by(animal_id) %>%
summarize(spl_type=length(unique(specimen_type))) %>%
filter(spl_type>1)
H5<-H4 %>%
filter(animal_id %in% Rmsp$animal_id & animal_id %in% unique(Rp$animal_id))
length(unique(H5$animal_id))## [1] 220then look at proportion of positive by sample types
H5%>%
group_by(specimen_type) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = round(ind_count_pos/ind_count*100,1)) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [8 x 5]
## Groups: <by row>
##
## # A tibble: 8 x 5
## specimen_type ind_count ind_count_pos percent_pos confint
## <fct> <int> <int> <dbl> <chr>
## 1 brain 16 5 31.2 12.1 - 58.5
## 2 feces 2 1 50 9.5 - 90.5
## 3 kidney 13 3 23.1 6.2 - 54
## 4 lung 51 27 52.9 38.6 - 66.8
## 5 oral swab 219 175 79.9 73.9 - 84.9
## 6 small intestine 155 80 51.6 43.5 - 59.7
## 7 spleen 1 1 100 5.5 - 100
## 8 urine/urogenital swab 1 0 0 0 - 94.5Paragraph 5 rodent farm results
Proportion of positive animals by genus in wildife farms
H4 %>%
filter(primary_interface_group=="wildlife farm") %>%
group_by(new_genus) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [4 x 5]
## Groups: <by row>
##
## # A tibble: 4 x 5
## new_genus ind_count ind_count_pos percent_pos confint
## <chr> <int> <int> <dbl> <chr>
## 1 Hystrix 331 20 6.04 3.8 - 9.3
## 2 Rattus 1 1 100 5.5 - 100
## 3 Rhizomys 96 6 6.25 2.6 - 13.6
## 4 Sciuridae 1 0 0 0 - 94.5overall, at site- and animal-level
H4 %>%
filter(primary_interface_group=="wildlife farm") %>%
group_by(state_prov) %>%
summarize(site_count=length(unique(site_name)),pos_sites=length(unique(site_name[confirmation_result=="Positive"])),siteprev=pos_sites/site_count*100,
ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))## Source: local data frame [1 x 8]
## Groups: <by row>
##
## # A tibble: 1 x 8
## state_prov site_count pos_sites siteprev ind_count ind_count_pos percent_pos
## <fct> <int> <int> <dbl> <int> <int> <dbl>
## 1 Dong Nai 28 17 60.7 429 27 6.29
## # … with 1 more variable: confint <chr>and disaggregated by genus and season (showing limitations in sampling design):
H4 %>%
filter(primary_interface_group=="wildlife farm") %>%
group_by(new_genus, hseason) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100)## # A tibble: 6 x 5
## # Groups: new_genus [4]
## new_genus hseason ind_count ind_count_pos percent_pos
## <chr> <chr> <int> <int> <dbl>
## 1 Hystrix Dry 296 20 6.76
## 2 Hystrix Wet 35 0 0
## 3 Rattus Wet 1 1 100
## 4 Rhizomys Dry 95 6 6.32
## 5 Rhizomys Wet 1 0 0
## 6 Sciuridae Dry 1 0 0and season only (aggregated on genus):
H4 %>%
filter(primary_interface_group=="wildlife farm") %>%
group_by(hseason) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
percent_pos = ind_count_pos/ind_count*100)## # A tibble: 2 x 4
## hseason ind_count ind_count_pos percent_pos
## <chr> <int> <int> <dbl>
## 1 Dry 392 26 6.63
## 2 Wet 37 1 2.70Paragraph 6 Bat findings
Subset the data for bats
H3<-subset(H1, taxa_group=="Bats")
length(unique(H3$animal_id)) #check## [1] 375Look at proportion positive per taxa based on animal ID
H3.1 <- H3 %>%
group_by(taxa_group, new_genus) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
ind_percent_pos = ind_count_pos/ind_count*100,
) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))
H3.1## Source: local data frame [3 x 6]
## Groups: <by row>
##
## # A tibble: 3 x 6
## taxa_group new_genus ind_count ind_count_pos ind_percent_pos confint
## <fct> <chr> <int> <int> <dbl> <chr>
## 1 Bats Cynopterus 2 0 0 0 - 80.2
## 2 Bats Microchiroptera 313 234 74.8 69.5 - 79.4
## 3 Bats Pteropus 60 4 6.67 2.2 - 17Seasonality for the Soc Trang bat pagoda
Look at proportion positive per season
H3.2 <- H3 %>%
filter(new_genus=="Pteropus") %>%
group_by(hseason) %>%
summarize(ind_count=length(unique(animal_id)),
ind_count_pos=length(unique(animal_id[confirmation_result=="Positive"])),
ind_percent_pos = ind_count_pos/ind_count*100,
) %>%
rowwise() %>%
mutate(confint=paste0(round(100*prop.test(ind_count_pos,ind_count, conf.level=0.95)$conf.int,1),collapse = ' - '))
H3.2## Source: local data frame [2 x 5]
## Groups: <by row>
##
## # A tibble: 2 x 5
## hseason ind_count ind_count_pos ind_percent_pos confint
## <chr> <int> <int> <dbl> <chr>
## 1 Dry 49 1 2.04 0.1 - 12.2
## 2 Wet 11 3 27.3 7.3 - 60.7Fisher exact test:
H1bp<-H1 %>%
filter(site_name=="Bat pagoda - Soc Trang", genus=="Pteropus") %>%
group_by(animal_id) %>%
summarize(coronavirus=length(unique(animal_id[confirmation_result=="Positive"])),
season = unique(hseason),
site_name = unique(site_name))
fisher.test(table(H1bp$season,H1bp$coronavirus))##
## Fisher's Exact Test for Count Data
##
## data: table(H1bp$season, H1bp$coronavirus)
## p-value = 0.01726
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
## 1.178107 956.801500
## sample estimates:
## odds ratio
## 16.63662This suggest an increased risk of being positive for cov during wet season for the Soc Trang pteropus
Phylogenetic analysis
List viral taxonomic units
H1$virus[H1$virus=="NULL"] <- NA
H1 %>%
distinct(virus)## virus
## 1 strain of Bat coronavirus 512/2005
## 2 PREDICT_CoV-35
## 3 <NA>
## 4 PREDICT_CoV-17
## 5 strain of Murine coronavirus
## 6 strain of Longquan Aa mouse coronavirus
## 7 strain of Infectious bronchitis virus (IBV)Number of viruses obtained with Watanbe protocol:
H1 %>%
filter(test_requested_protocol=='Modified Watanabe et al, RdRp gene') %>%
distinct(virus)## virus
## 1 strain of Bat coronavirus 512/2005
## 2 <NA>
## 3 PREDICT_CoV-35
## 4 PREDICT_CoV-17
## 5 strain of Murine coronavirus
## 6 strain of Longquan Aa mouse coronavirusNumber of viruses obtained with Quan protocol:
H1 %>%
filter( test_requested_protocol=='Quan et al, RdRp gene')%>%
distinct(virus)## virus
## 1 strain of Bat coronavirus 512/2005
## 2 PREDICT_CoV-35
## 3 <NA>
## 4 strain of Murine coronavirus
## 5 strain of Longquan Aa mouse coronavirus
## 6 PREDICT_CoV-17
## 7 strain of Infectious bronchitis virus (IBV)allpos<-H1 %>%
filter(confirmation_result=="Positive") %>%
group_by(animal_id) %>%
summarize(ind_count=length(unique(animal_id)),
want_pos=length(unique(animal_id[test_requested_protocol=="Modified Watanabe et al, RdRp gene" & confirmation_result=="Positive"])),
quan_pos=length(unique(animal_id[test_requested_protocol=="Quan et al, RdRp gene" & confirmation_result=="Positive"])),
both_tests_pos = ifelse(want_pos+quan_pos==2,1,0))
colSums(allpos[,2:5])## ind_count want_pos quan_pos both_tests_pos
## 504 433 410 339Viruses found in bats
H1 %>%
filter(virus_group=="PREDICT_CoV-17") %>%
group_by(new_genus) %>%
summarize(n_Infected_with_PCov17=length(unique(animal_id)))## # A tibble: 2 x 2
## new_genus n_Infected_with_PCov17
## <chr> <int>
## 1 Microchiroptera 1
## 2 Pteropus 3H1 %>%
filter(virus_group=="PREDICT_CoV-35") %>%
group_by(new_genus) %>%
summarize(n_Infected_with_PCov35=length(unique(animal_id)))## # A tibble: 2 x 2
## new_genus n_Infected_with_PCov35
## <chr> <int>
## 1 Microchiroptera 38
## 2 Pteropus 1H1 %>%
filter(virus_group=="Bat coronavirus 512/2005") %>%
group_by(new_genus) %>%
summarize(n_Infected_with_BCov5122005=length(unique(animal_id)))## # A tibble: 4 x 2
## new_genus n_Infected_with_BCov5122005
## <chr> <int>
## 1 Hystrix 19
## 2 Microchiroptera 216
## 3 Rattus 1
## 4 Rhizomys 5See also Table 1.
Viruses found in rodents
H1 %>%
group_by(taxa_group, new_genus, virus_group) %>%
summarize(ind_count_virus=length(unique(animal_id))) %>%
mutate(virus_ind_count=paste0(virus_group, " (", ind_count_virus,")"))%>%
group_by(taxa_group, new_genus) %>%
summarize(viral_species=paste0(virus_ind_count ,collapse = ', ')) %>%
arrange(desc(taxa_group))## # A tibble: 8 x 3
## # Groups: taxa_group [2]
## taxa_group new_genus viral_species
## <fct> <chr> <chr>
## 1 Rodents & Shr… Field Rats Longquan Aa mouse coronavirus (56), Murine corona…
## 2 Rodents & Shr… Hystrix Bat coronavirus 512/2005 (19), Infectious bronchi…
## 3 Rodents & Shr… Rattus Bat coronavirus 512/2005 (1), NULL (1)
## 4 Rodents & Shr… Rhizomys Bat coronavirus 512/2005 (5), Infectious bronchit…
## 5 Rodents & Shr… Sciuridae NULL (1)
## 6 Bats Cynopterus NULL (2)
## 7 Bats Microchirop… Bat coronavirus 512/2005 (216), NULL (126), PREDI…
## 8 Bats Pteropus NULL (59), PREDICT_CoV-17 (3), PREDICT_CoV-35 (1)and see code for Table 1 above
Details on IBV detection:
H1 %>%
filter(virus_group=="Infectious bronchitis virus (IBV)") %>%
group_by(primary_interface_group,taxa_group,species_scientific_name, site_name, id_certainty) %>%
summarize(ind_count=length(unique(animal_id)))## # A tibble: 2 x 6
## # Groups: primary_interface_group, taxa_group, species_scientific_name,
## # site_name [2]
## primary_interfac… taxa_group species_scienti… site_name id_certainty ind_count
## <fct> <fct> <fct> <fct> <fct> <int>
## 1 wildlife farm Rodents &… Hystrix brachyu… Farm 25 … field ID ce… 1
## 2 wildlife farm Rodents &… Rhizomys sp. Farm 4 -… field ID ce… 1