Case Study: Cheese

library(MolPad)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

Load Data

In the process of producing cheese, regular washing with a brine solution is an aging practice that can homogenize the microbial communities on the cheese’ surface and facilitate intermicrobial interactions. In the following parts, we will analyze a longitudinal data set of three washed-rind cheese communities collected during cheese ripening.

data("cheese")

Preprocess

Data and Annotations

Here we select a ‘type’ column with phylum to describe the characteristic of cheese data. Also, the columns phylum and class are taken as tags for elemental composition.

cheesedata <- cheese |> 
         select(ID, A_1:C_5, phylum) |> 
         rename(type=phylum) |> 
            pre_process()

pathchee <- gAnnotation(annotations,"phylum","class")

After pre-processing, here is what the two datasets that we will put into the dashboard look like:

cheesedata[112:115,]
#> # A tibble: 4 × 11
#>   ID      A_1    A_2    A_3    A_4    A_5    C_1    C_3    C_4    C_5 type 
#>   <chr> <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl> <chr>
#> 1 112   0.943 -0.471 -0.471 -0.471 -0.471  2.36  -0.471 -0.471 -0.471 Other
#> 2 113   0.786 -0.124  0.126 -0.623 -0.637  2.33  -0.667 -0.593 -0.593 Other
#> 3 114   1.22  -1.09  -0.430 -0.829 -0.170  1.43  -1.30   0.455  0.715 Other
#> 4 115   2.67  -0.333 -0.333 -0.333 -0.333 -0.333 -0.333 -0.333 -0.333 Other
pathchee[112:115,]
#> # A tibble: 4 × 9
#>   ID    GO_ID KEGG_ID          domain Pathway taxonomic.scope order family genus
#>   <chr> <chr> <chr>            <chr>  <chr>   <chr>           <chr> <chr>  <chr>
#> 1 145   -     ko:K00004        Bacte… Pseudo… Unknown         NA    NA     NA   
#> 2 147   -     ko:K00004        Bacte… Actino… Actinomycetes   NA    NA     NA   
#> 3 148   -     ko:K00004        Bacte… Bacill… Bacilli         Baci… Bacil… Lysi…
#> 4 149   -     ko:K00004,ko:K0… Bacte… Actino… Actinomycetes   NA    NA     NA

Cluster Input

Then we generate clusters from the long dataset above.

cluschee <- gClusters(cheesedata,ncluster = 10,elbow.max=15)

network input

After that, we got the network result from the clusters. We pick top 3 predictors for each feature, and save the network output for future use.

networkchee <- gNetwork(cluschee,ntop = 3)

View more details of the results with the following function:

gNetwork_view(networkchee)

Run Dashboard

Declare the annotations, and then run the dashboard.

gDashboard(cheesedata,
           cluschee,
           pathchee,
           networkchee,
           id_colname = c("GO_ID","KEGG_ID"),
           id_type = c("GO","KEGG"))