Project notebook with R¶

Easy access to all the experimental data, statistical analysis results and visualization of reports using R
All the R functionality at hand

Import R libraries¶

Make sure that all the libraries you plan to use have been previously installed.

If not, you can install them through the notebook.

E.g. install.packages(“tidyverse”)

[10]:

library(rhdf5)
library(rjson)
library(foreach)
library(plotly)
library(tidyverse)
library(networkD3)

Define custom functions¶

Within the Jupyter notebook you can create all the functions relevant to you work that are not implemented in the analytics core. This is the case of R functions. Here we define R functions that allow us to load a project from the CKG database, access different datasets and plots, and visualize the latter in the notebook.

[11]:

importGraph <- function(json_network){
  edges_df = data.frame(matrix(unlist(json_network$net_json$links), nrow=length(json_network$net_json$links), byrow=T))
  colnames(edges_df)<-names(unlist(json_network$net_json$links[[1]]))
  edges_df <- edges_df[c("source", "target", "width")]
  colnames(edges_df)<-c("source", "target", "weight")
  nodes = data.frame(matrix(unlist(json_network$net_json$nodes), nrow=length(json_network$net_json$nodes), byrow=T))
  colnames(nodes)<-names(unlist(json_network$net_json$nodes[[1]]))
  nodes['ident']<- 0:(dim(nodes)[1]-1)
  nodes <- nodes[c("ident", "id", "color")]
  colnames(nodes) <- c("id", "label", "color")
  edges <- edges_df %>%
    left_join(nodes, by = c("source" = "label")) %>%
    rename(from = id)
  edges <- edges %>%
    left_join(nodes, by = c("target" = "label")) %>%
    rename(to = id)
  edges <- edges[c("from", "to", "weight")]
  library(networkD3)
  g<-forceNetwork(Links = edges, Nodes = nodes,
                  height = 850, width = 800,
                  Source = "from", Target = "to",
                  Value = "weight", NodeID = "label",
                  Group = "color", opacity = 0.8,opacityNoHover = 1,
                  zoom = T, linkWidth = JS("function(d) { return Math.sqrt(d.weight); }"))
  return(g)
}

read_report <- function(report_file, report_name){
  plots <- c()
  nets <- c()
  if(file.exists(report_file)){
    content <- h5ls(report_file)
    foreach(group=unique(content$group)) %do% {
      if(group != "/"){
        if(!grepl('Table', group)){
          group = substring(group, 2)
          report_figure <- h5read(report_file, group)
          foreach(name=unique(names(report_figure))) %do% {
            if(grepl('figure',name)){
              report_figure_str<-fromJSON(report_figure[[name]][1])
              p = report_figure_str$props$figure
              plot <- list(plot=p)
              names(plot) <- paste(group,name,sep="_")
              plots<-append(plots, plot)
            }
            else if(grepl('net', name)){
              report_net_str <- fromJSON(report_figure[[name]][1])
              net <- list(net=report_net_str)
              names(net) <- name
              nets<-append(nets, net)
            }
          }
        }
      }
    }
  }
  h5closeAll()
  report <- list(name = report_name, plots = plots, nets = nets)
  class(report) <- "report"

  return(report)
}

read_dataset <- function(dataset_file, dataset_name){
  dataframes <- c()
  if(file.exists(dataset_file)){
    content <- h5ls(dataset_file)
    foreach(group=unique(content$group)) %do% {
      if(group != "/"){
        group = substring(group, 2)
        dataframe_json <- h5read(dataset_file, group)
        foreach(name=unique(names(dataframe_json))) %do% {
          if(is.null(names(dataframe_json[[name]]))){
          dataframe_str<-fromJSON(dataframe_json[[name]][1])}
          df <- data.frame(matrix(unlist(dataframe_str), nrow=length(dataframe_str), byrow=T))
          colnames(df)<-names(unlist(dataframe_str[[1]]))
          dataframe <- list(data=df)
          names(dataframe) <- name
          dataframes<-append(dataframes, dataframe)
        }
      }
    }
  }
  h5closeAll()
  dataset <- list(name=dataset_name, data=dataframes)
  class(dataset) <- "dataset"

  return(dataset)
}


load_project <- function(project_id, dataset){
  plots = c()
  nets = c()
  project_report_dir <- paste("../../../data/reports/", project_id, sep = "")
  project_dataset_report_dir <- paste(project_report_dir,dataset, sep="/")
  report_file = paste(project_dataset_report_dir, "report.h5", sep = "/")
  datasets_file = paste(project_dataset_report_dir, paste(dataset,"dataset.h5",sep='_'), sep = "/")
  report = read_report(report_file, paste(project_id, "Report", dataset))
  datasets = read_dataset(datasets_file, paste(project_id, "Dataset", dataset))
  project <- list(id=project_id, report=report, datasets=datasets)
  return(project)
}

Create a project object by loading an existent report¶

[12]:

p = load_project("P0000001", "proteomics")

Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [56749] is not a sub-multiple or multiple of the number of rows [17298]”
Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [942] is not a sub-multiple or multiple of the number of rows [188]”
Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [140] is not a sub-multiple or multiple of the number of rows [8]”
Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [152735] is not a sub-multiple or multiple of the number of rows [26135]”
Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [152735] is not a sub-multiple or multiple of the number of rows [26135]”
Warning message in matrix(unlist(dataframe_str), nrow = length(dataframe_str), byrow = T):
“data length [19912] is not a sub-multiple or multiple of the number of rows [6187]”

Visualize the list of plots contained in the report¶

[13]:

names(p$report$plots)

'0~proteomics_pipeline~cytoscape_network_0_figure'
'11~stratification_description~description_0_figure'
'12~stratification_pca~pca_0_figure'
'13~regulation_description~description_0_figure'
'15~regulation_samr~volcanoplot_0_figure'
'15~regulation_samr~volcanoplot_1_figure'
'15~regulation_samr~volcanoplot_2_figure'
'15~regulation_samr~volcanoplot_3_figure'
'15~regulation_samr~volcanoplot_4_figure'
'15~regulation_samr~volcanoplot_5_figure'
'15~regulation_samr~volcanoplot_6_figure'
'15~regulation_samr~volcanoplot_7_figure'
'15~regulation_samr~volcanoplot_8_figure'
'15~regulation_samr~volcanoplot_9_figure'
'22~literature_associations_publications_abstracts~wordcloud_0_figure'
'2~peptides~barplot_0_figure'
'4~proteins~barplot_0_figure'
'6~modifications~facetplot_0_figure'
'8~coefficient_variation_coefficient_of_variation~scatterplot_matrix_0_figure'
'9~ranking_ranking_with_markers~ranking_0_figure'

Access a specific plot and use plotly to visualize it¶

[18]:

plotly_build(p$report$plots$`2~peptides~barplot_0_figure`)

In the case of networks, they have to be converted from a json format, to an edge list. We use R’s networkd3 to create a D3 JavaScript force directed network graph from it.¶

[15]:

importGraph(p$report$nets$`0_net`)

Warning message:
“Column `source`/`label` joining factors with different levels, coercing to character vector”
Warning message:
“Column `target`/`label` joining factors with different levels, coercing to character vector”

We can also, easily, access the different datasets from the project¶

[16]:

names(p$datasets$data)

'complex_associations'
'correlation_correlation'
'disease_associations'
'drug_associations'
'go annotation'
'go_enrichment_Biological_processes_regulation_enrichment'
'interaction_network'
'literature_associations_publications_abstracts'
'number of modified proteins'
'number of peptides'
'number of proteins'
'original'
'overview statistics_summary'
'pathway annotation'
'pathway_enrichment_Pathways_regulation_enrichment'
'processed'
'protein biomarkers'
'regulated'
'regulation table'
'Data Matrix Shape'
'Stats'

[17]:

p$datasets$data$correlation_correlation

A data.frame: 74922 × 6
node1	node2	weight	pvalue	padj	rejected
<fct>	<fct>	<fct>	<fct>	<fct>	<fct>
A30~A2MYE2	A2M~P01023	0.2892365267	0.0	0.0	TRUE
ABI3BP~Q7Z7G0	A30~A2MYE2	-0.3129542643	0.0	0.0	TRUE
ACE~P12821	A2M~P01023	0.2904851136	0.00059052	0.00164552	TRUE
ACE~P12821	ABI3BP~Q7Z7G0	0.0026333708	3e-08	2e-07	TRUE
ACTB~P60709	A2M~P01023	-0.1503252993	1.4e-07	8.5e-07	TRUE
ACTB~P60709	ABI3BP~Q7Z7G0	-0.1213572078	0.00959561	0.01965832	TRUE
ACTB~P60709	ACE~P12821	-0.0120569051	0.00172883	0.00428982	TRUE
ACTN1~P12814	A30~A2MYE2	-0.0953185874	4.137e-05	0.00014992	TRUE
ACTN1~P12814	ABI3BP~Q7Z7G0	0.2159122385	0.0	0.0	TRUE
ACTN1~P12814	ACE~P12821	0.0337919091	0.00146044	0.00369332	TRUE
ACTN1~P12814	ACTB~P60709	0.2556916718	2.296e-05	8.778e-05	TRUE
ADA2~Q9NZK5	A2M~P01023	0.2405633762	0.0	1e-08	TRUE
ADA2~Q9NZK5	ABI3BP~Q7Z7G0	-0.1378733998	0.0068122	0.01452743	TRUE
ADA2~Q9NZK5	ACTB~P60709	0.0488541846	0.00228905	0.00551185	TRUE
ADA2~Q9NZK5	ACTN1~P12814	0.1606489207	9.2e-07	4.71e-06	TRUE
ADAMTS13~Q76LX8	A2M~P01023	0.228562439	1.3e-07	7.7e-07	TRUE
ADAMTS13~Q76LX8	ACE~P12821	0.1100432326	9.62e-06	3.996e-05	TRUE
ADAMTS13~Q76LX8	ACTB~P60709	-0.1298491445	0.0	0.0	TRUE
ADAMTS13~Q76LX8	ADA2~Q9NZK5	0.0882539027	0.0	0.0	TRUE
ADAMTSL4~Q6UY14	A2M~P01023	-0.0167111069	0.0258627	0.0469946	TRUE
ADAMTSL4~Q6UY14	ABI3BP~Q7Z7G0	-0.0297828179	2.1e-07	1.21e-06	TRUE
ADAMTSL4~Q6UY14	ACE~P12821	0.1867241047	2e-08	1.1e-07	TRUE
ADAMTSL4~Q6UY14	ACTB~P60709	0.0422307278	0.0	0.0	TRUE
ADAMTSL4~Q6UY14	ACTN1~P12814	0.167608862	0.0	0.0	TRUE
ADAMTSL4~Q6UY14	ADA2~Q9NZK5	-0.0456140059	3.96e-06	1.779e-05	TRUE
ADH4~P08319	A2M~P01023	-0.1135971714	0.01333768	0.02626505	TRUE
ADH4~P08319	A30~A2MYE2	0.0504134162	0.0	0.0	TRUE
ADH4~P08319	ACE~P12821	0.065413239	0.0	0.0	TRUE
ADH4~P08319	ACTB~P60709	-0.0805120054	0.0	0.0	TRUE
ADH4~P08319	ACTN1~P12814	0.0217710049	0.0	0.0	TRUE
⋮	⋮	⋮	⋮	⋮	⋮
scFv~Q65ZC9	SERPINA5~P05154	-0.113307701	1.005e-05	4.157e-05	TRUE
scFv~Q65ZC9	SERPINA7~P05543	0.0029215059	0.00655886	0.01404733	TRUE
scFv~Q65ZC9	SERPINF1~P36955	-0.0358642698	0.0	0.0	TRUE
scFv~Q65ZC9	SERPING1~P05155	-0.4645053826	0.01186057	0.02368813	TRUE
scFv~Q65ZC9	SHBG~P04278	0.0597719401	5.92e-06	2.568e-05	TRUE
scFv~Q65ZC9	SNCA~P37840	0.0526730489	0.0	0.0	TRUE
scFv~Q65ZC9	SOD3~P08294	0.065714844	0.00134567	0.00343099	TRUE
scFv~Q65ZC9	SPARCL1~Q14515	0.2050409349	4e-08	2.6e-07	TRUE
scFv~Q65ZC9	SPINK5~Q9NQ38	0.0993925633	0.0	0.0	TRUE
scFv~Q65ZC9	SPP2~Q13103	0.1951832721	0.00019287	0.00060287	TRUE
scFv~Q65ZC9	STMN1~P16949	-0.1916459118	0.0	0.0	TRUE
scFv~Q65ZC9	THBS1~P07996	-0.0833391798	1.277e-05	5.16e-05	TRUE
scFv~Q65ZC9	THBS4~P35443	-0.010593396	0.00043177	0.00124437	TRUE
scFv~Q65ZC9	TKT~P29401	-0.244463802	0.00138887	0.00353092	TRUE
scFv~Q65ZC9	TNC~A0A024R884	0.2552310889	0.0	0.0	TRUE
scFv~Q65ZC9	TNN~Q9UQP3	-0.1075661031	0.00242762	0.00580852	TRUE
scFv~Q65ZC9	TNXB~P22105	0.2817641092	3.13e-05	0.00011634	TRUE
scFv~Q65ZC9	TPI1~P60174	-0.2298225046	0.0	1e-08	TRUE
scFv~Q65ZC9	V1-13~Q5NV69	0.1054214299	0.00306398	0.00714884	TRUE
scFv~Q65ZC9	V2-13~Q5NV73	0.4771154135	2e-08	1.3e-07	TRUE
scFv~Q65ZC9	V4-2~Q5NV82	0.2165834919	1.139e-05	4.654e-05	TRUE
scFv~Q65ZC9	V5-2~A2MYC8	0.1808639979	0.0	0.0	TRUE
scFv~Q65ZC9	V5-4~Q5NV79	-0.0853365176	6.394e-05	0.0002221	TRUE
scFv~Q65ZC9	VASN~Q6EMK4	0.1403713766	4e-08	2.9e-07	TRUE
scFv~Q65ZC9	VCAM1~P19320	-0.0542730041	0.00915923	0.01886712	TRUE
scFv~Q65ZC9	VH6DJ~A2N0T4	0.2625776727	8e-08	5e-07	TRUE
scFv~Q65ZC9	VIM~P08670	0.0453630504	0.00012994	0.00042167	TRUE
scFv~Q65ZC9	VK3~A2N2F4	0.1745551542	0.02251528	0.04162447	TRUE
scFv~Q65ZC9	VNN1~O95497	-0.2329496712	0.01174417	0.02348694	TRUE
scFv~Q65ZC9	YWHAZ~P63104	0.2255080844	2.1e-07	1.21e-06	TRUE

[ ]: