"GA3 Workflow.zip" consists of phenotypic extraction files
"Tutorial.zip" consists of txt. file explaining how to run both source data and user generated input data
"NMP Source Data.zip" consists of folder with all raw data and scripts to run for graphing output
Below is explaination of code to users
>library(umap)
>library(ggplot2)
>library(factoextra)
>library("writexl")
Import functions for data and graphing
(umap) = R studio package UMAP (Version 0.2.10.0) - for dimensional reduction
(ggplot2) = R studio package ggplot2 (version 3.3. 4) - for plotting
(factoextra) = R studio package factoextra (Version 1.0.7) - for clustering
("writexl") = R studio package UMAP (Version 1.5.0) - for excel exporting
>Reduced_file_name <-file name
Imported file name on the right and reduce file name on left
>Name_data <-reduced file name[ ,columns]
On the right put reduced file name and in brackets leave space empty for rows and after comma choose columns from file that will be used as 4 parameters. On left name data.
>Name_scale <-scale(Name_data)
Scale the data set.
>umap_result <-umap(Name_scale, n_neighbors= #, n_components = #)
Apply dimension reduction. Neighbors # is traditionally set to 20. Components # sets dimensionality. Component of 1 determines scores.
>df_con <-data.frame(umap_result[["layout"]])
Frame all the data created by the UMAP function into an exportable format.
>x <-df_con[ ,1]
>y <-df_con [ ,2]
>x <-unlist(x)
>y <-unlist(y)ggplot(df_con, aes(x=x, y=y))+ geom_point()
Parse out data points then plot for raw UMAP.
>km.res <-kmeans(T_scale, #, nstart = #)
>fviz_cluster(km.res, df_con, frame = FALSE, geom = "point")
NMP Code.docx
Apply clustering. # determines number of clusters. nStart # is traditionally set to 5.
>km <-unlist(km.res)
>km.matrix <-as.matrix(km)
>km.df <-data.frame(km.matrix)
Frame clustering data into an exportable format.
>write_xlsx(km.df, "C:Users/Desktop\
FILE_NAME.xlsx")
>write_xlsx(df_con, "C:Users/Desktop\
FILE_NAME.xlsx")
Export data into an exportable excel.