EcoPlate解析 (2)

#We need to separate the data set into each sample with multiple measurements

class(meta_test$sample_ID[3]

#number of samples

max(meta_test$sample_ID)

#OR

length(levels(meta_test$date_time_sample))

#trial to generate the list of list

test_list<-list()

test_list[[1]]<-list()

test_list[[1]][[1]]<-osaka_data[[1]]

test_list[[1]][[2]]<-osaka_data[[2]]

test_list[[1]]

test_list[[1]][[2]]

#separating the data to sample ID-specific list

sample_list<-list()

no_sample<-max(meta_test$sample_ID)

for(i in 1:no_sample) {

  sample_list[[i]]<-list()

}

#summarizing the metadata into sample-ID-specific list

metadata_list<-list()

no_sample<-max(meta_test$sample_ID)

for(i in 1:no_sample) {

  metadata_list[[i]]<-as.data.frame(meta_test[1,])

}

#Check the number of repeated measurements

sum(meta_test$sample_ID==5)

i<-1 #initialization

for(j in 1: no_sample) {

  no_measurement<-sum(meta_test$sample_ID==j) #check the number of repeated measurements

  #loop for loading all measurement results from sample index j

  for(k in 1:no_measurement) {

    sample_list[[j]][[k]]<-osaka_data[[i]] #copy the data of each measurement into sample_list[[sample_index]]

    metadata_list[[j]][k,]<-meta_test[i,] #copy the metadata

    i<-i+1 #update the index of the raw data (list osaka_data)

  }

}

#example

sample_id<-3

sample_list[[sample_id]]

metadata_list[[sample_id]]

##########[2] Definitions of functions to data arrangement (integration, maximum, average, minimum, etc)##################

#The function for taking integration (=calculation of the area below the color development curve) even when there are gaps in measurement dates

##########################list of parameters

#ecop_series: data source of ecopalte color patterns from the single sample ID (list format)

#metadata: metadata, we need the information of measured time

#min_interval_hour, the minimum interval hour, here 24 hours

##################Output is a dataframe

integ_ecoplate <- function(ecop_series, metadata, min_interval_hour=24)

{

 data_eco <- list() #prepare an empty list

 sampled_index <- metadata$measured_time/min_interval_hour   #convert hours to n-th observation (n-th day) 

 end_sample <- max(sampled_index)   #The final date of the measurement

 length_period <- length(ecop_series) #size of the samples, generally less or equal to end_sample, because the interval of the observation is ofen longer than min_interval_hour

for(i in 1: length_period) data_eco[[sampled_index[i]]] <-ecop_series[[i]] #load the ecoplate data only for sampled dates, some of the element of data_eco can be empty (NULL)

data_integ <- data.frame()  #prepare a data frame

data_integ <- data_eco[[1]] #Note that "data_eco" should be a list; each item is the information of color development from each measurement date.

for(i in 2:end_sample) {

    #first check the NULL dates (lack in measurement)

    if(is.null(data_eco[[i]])) {

      for(j in (i+1):end_sample) { #search for all NULL dates until non-Null date

        if(!is.null(data_eco[[j]])) {  #when encountring the non-Null date (j)

          for(k in i: (j-1)) data_eco[[k]] = ((j - k)*data_eco[[i-1]] + (k - i + 1)*data_eco[[j]])/(j - i + 1)  #use linear line of two non-NULL dates (i-1 & j)

          break   #skip further loop when non-zero data is found

          }#end of if data_eco[[j]]

      }#end of for j

    }#end of if data_eco[[i]]  

data_integ <- data_integ + data_eco[[i]]  #each element of data_eco[[i]] was individually added.

  }#end of for loop i

data_integ/(end_sample*min_interval_hour)  #output, take the average value over the integration period, i.e., the normalization by integration period

}

> sample_id<-4

> integ_ecoplate(ecop_series=sample_list[[sample_id]], metadata_list[[sample_id]], min_interval_hour=24)

           V1          V2          V3          V4          V5          V6          V7          V8          V9         V10         V11         V12

A 0.003101190 0.003029762 0.005258929 0.009366071 0.013357143 0.003461310 0.006866071 0.012779762 0.002916667 0.004008929 0.005410714 0.009952381

B 0.013651786 0.013860119 0.008663690 0.013125000 0.014080357 0.032985119 0.008386905 0.024922619 0.013059524 0.014000000 0.009032738 0.021288690

C 0.013592262 0.003193452 0.002702381 0.004514881 0.014809524 0.003488095 0.002398810 0.003205357 0.012190476 0.003354167 0.002455357 0.003413690

D 0.009056548 0.005023810 0.006800595 0.016619048 0.008199405 0.004750000 0.005482143 0.012154762 0.007809524 0.007529762 0.015559524 0.016934524

E 0.014101190 0.012833333 0.004172619 0.003681548 0.014827381 0.015178571 0.003818452 0.003779762 0.013169643 0.017023810 0.003514881 0.004056548

F 0.022017857 0.004964286 0.005997024 0.010467262 0.027886905 0.003592262 0.003937500 0.008949405 0.017943452 0.003363095 0.004735119 0.014077381

H 0.004181548 0.003544643 0.004979167 0.004633929 0.003872024 0.003217262 0.005169643 0.006580357 0.003767857 0.003535714 0.005550595 0.004443452

G 0.003776786 0.005199405 0.005005952 0.009931548 0.003166667 0.005229167 0.004812500 0.006276786 0.002913690 0.005029762 0.003267857 0.011389881

> 

#integrating all samples

integrated_osaka<-list() #prepare emplty list

no_sample<-max(meta_test$sample_ID)

for(i in 1:no_sample) integrated_osaka[[i]] <- integ_ecoplate(ecop_series=sample_list[[i]], metadata_list[[i]], min_interval_hour=24)

integrated_osaka

#extracting only the final day measurement data

#example

sample_list[[1]][[length(sample_list[[1]])]]

finalday_osaka<-list() #prepare emplty list

no_sample<-max(meta_test$sample_ID)

for(i in 1:no_sample) finalday_osaka[[i]] <-sample_list[[i]][[length(sample_list[[i]])]]

finalday_osaka

data_ave1<-(data_f$V1+data_f$V5+data_f$V9)/3.0  

#The function for averaging triplicate

#The parameter is data frame

#Output is a vector

ave_ecoplate <- function(data_f){

  data_ave1<-(data_f$V1+data_f$V5+data_f$V9)/3.0  #take average

  data_ave2<-(data_f$V2+data_f$V6+data_f$V10)/3.0

  data_ave3<-(data_f$V3+data_f$V7+data_f$V11)/3.0

  data_ave4<-(data_f$V4+data_f$V8+data_f$V12)/3.0

  data_sum<-append(append(append(data_ave1,data_ave2),data_ave3),data_ave4)  #combine data

  data_sum_nor<-data_sum - data_sum[1] #normalizing by water well

  data_sum_nor #output

}

> sample_id<-2

> ave_ecoplate(finalday_osaka[[sample_id]])

 [1]  0.000000000  0.466666667  0.688333333  0.035000000  0.837666667  0.087666667  0.541333333 -0.333000000 -0.191666667 -0.324333333 -0.346333333 -0.072666667

[13]  0.784666667 -0.161333333 -0.342000000 -0.174333333  0.312666667  0.323333333 -0.372666667 -0.268666667 -0.314333333 -0.009666667 -0.350333333 -0.305000000

[25] -0.114333333  0.845333333 -0.329000000  0.524000000 -0.327333333  0.125666667 -0.269333333 -0.090333333

#The function for taking maximum of triplicate

#The parameter is data frame

#Output is a vector

max_ecoplate <- function(data_f){

  data_max<-max(data_f$V1[1],data_f$V5[1], data_f$V9[1])

  for(i in 2:8) data_max <-append(data_max, max(data_f$V1[i], data_f$V5[i], data_f$V9[i]))

  for(i in 1:8) data_max <-append(data_max, max(data_f$V2[i], data_f$V6[i], data_f$V10[i]))

  for(i in 1:8) data_max <-append(data_max, max(data_f$V3[i], data_f$V7[i], data_f$V11[i]))

  for(i in 1:8) data_max <-append(data_max, max(data_f$V4[i], data_f$V8[i], data_f$V12[i]))

  data_max_nor <- data_max - data_max[1]  #normalizing by water well

  data_max_nor  #output

}

#The function for takign minumu of triplicate

#The parameter is data frame

#Output is a vector

min_ecoplate <- function(data_f){

  data_min<-min(data_f$V1[1], data_f$V5[1], data_f$V9[1])

  for(i in 2:8) data_min <-append(data_min, min(data_f$V1[i], data_f$V5[i], data_f$V9[i]))

  for(i in 1:8) data_min <-append(data_min, min(data_f$V2[i], data_f$V6[i], data_f$V10[i]))

  for(i in 1:8) data_min <-append(data_min, min(data_f$V3[i], data_f$V7[i], data_f$V11[i]))

  for(i in 1:8) data_min <-append(data_min, min(data_f$V4[i], data_f$V8[i], data_f$V12[i]))

  data_min_nor <- data_min - data_min[1]  #normalizing by water well

  data_min_nor  #output

}

#Integration + average

osaka_summary_integ_ave<-ave_ecoplate(integrated_osaka[[1]])

for(i in 2: length(integrated_osaka)) osaka_summary_integ_ave<-rbind(osaka_summary_integ_ave,ave_ecoplate(integrated_osaka[[i]]))

View(osaka_summary_integ_ave)

#Integration + max

osaka_summary_integ_max<-max_ecoplate(integrated_osaka[[1]])

for(i in 2: length(integrated_osaka)) osaka_summary_integ_max<-rbind(osaka_summary_integ_max,max_ecoplate(integrated_osaka[[i]]))

#Integration + min

osaka_summary_integ_min<-min_ecoplate(integrated_osaka[[1]])

for(i in 2: length(integrated_osaka)) osaka_summary_integ_min<-rbind(osaka_summary_integ_min,min_ecoplate(integrated_osaka[[i]]))

#finalday + average

osaka_summary_final_ave<-ave_ecoplate(finalday_osaka[[1]])

for(i in 2: length(finalday_osaka)) osaka_summary_final_ave<-rbind(osaka_summary_final_ave,ave_ecoplate(finalday_osaka[[i]]))

#finalday + max

osaka_summary_final_max<-max_ecoplate(finalday_osaka[[1]])

for(i in 2: length(finalday_osaka)) osaka_summary_final_max<-rbind(osaka_summary_final_max,max_ecoplate(finalday_osaka[[i]]))

#finalday + min

osaka_summary_final_min<-min_ecoplate(finalday_osaka[[1]])

for(i in 2: length(finalday_osaka)) osaka_summary_final_min<-rbind(osaka_summary_final_min,min_ecoplate(finalday_osaka[[i]]))

#convert them into dataframe, delete the control value (0)

class(osaka_summary_final_ave)

osaka_summary_integ_ave<-as.data.frame(osaka_summary_integ_ave[,-1])

osaka_summary_integ_max<-as.data.frame(osaka_summary_integ_max[,-1])

osaka_summary_integ_min<-as.data.frame(osaka_summary_integ_min[,-1])

osaka_summary_final_ave<-as.data.frame(osaka_summary_final_ave[,-1])

osaka_summary_final_max<-as.data.frame(osaka_summary_final_max[,-1])

osaka_summary_final_min<-as.data.frame(osaka_summary_final_min[,-1])

#prepare the meta data file again for sample_ID

metadata_osaka<-metadata_list[[1]][1,1:3]

for(i in 1:length(metadata_list)) metadata_osaka<-rbind(metadata_osaka,metadata_list[[i]][1,1:3])

metadata_osaka<-as.data.frame(metadata_osaka)

#write the result into csv files

write.csv(metadata_osaka, "./metadata_osaka.csv", row.names=F, quote=F)

write.csv(osaka_summary_integ_ave, "./osaka_summary_integ_ave.csv", row.names=F, quote=F)

write.csv(osaka_summary_integ_max, "./osaka_summary_integ_max.csv", row.names=F, quote=F)

write.csv(osaka_summary_integ_min, "./osaka_summary_integ_min.csv", row.names=F, quote=F)

write.csv(osaka_summary_final_ave, "./osaka_summary_final_ave.csv", row.names=F, quote=F)

write.csv(osaka_summary_final_max, "./osaka_summary_final_max.csv", row.names=F, quote=F)

write.csv(osaka_summary_final_min, "./osaka_summary_final_min.csv", row.names=F, quote=F)