Click here for page contents...

Prerequisites

It is recommended that you familiarise yourself with R first by sitting our Introduction to R tutorial.

It also requires that you have the DataSHIELD training environment installed on your machine, see our Installation Instructions for Linux, Windows, or Mac.

Help

DataSHIELD support is freely available in the DataSHIELD forum by the DataSHIELD community. Please use this as the first port of call for any problems you may be having, it is monitored closely for new threads.

DataSHIELD bespoke user support and also user training classes are offered on a fee-paying basis. Please enquire at datashield@newcastle.ac.uk for current prices.

Introduction

This is the second in a 6-part DataSHIELD tutorial series.

The other parts in this DataSHIELD tutorial series are:

1: Introduction and logging in
2: Basic statistics and data manipulations
3: Assign functions and tables
4: Plotting graphs
5: Subsetting
6: Modelling

Quick reminder for logging in:

Click here to expand...

Start R/RStudio

Load Packages

#load libraries
library(DSI)
library(DSOpal)
library(dsBaseClient)

Build your login dataframe

Build your login dataframe

builder <- DSI::newDSLoginBuilder()
builder <- DSI::newDSLoginBuilder()
builder$append(server = "server1", url = "https://opal-demo.obiba.org/",
user = "dsuser", password = "P@ssw0rd", driver = "OpalDriver", options='list(ssl_verifyhost=0, ssl_verifypeer=0)')
builder$append(server = "server2", url = "https://opal-demo.obiba.org/",
user = "dsuser", password = "P@ssw0rd", driver = "OpalDriver", options='list(ssl_verifyhost=0, ssl_verifypeer=0)')

logindata <- builder$build()

logindata <- builder$build()connections <- DSI::datashield.login(logins = logindata, assign = TRUE)
DSI::datashield.assign.table(conns = connections, symbol = "DST", table = c("CNSIM.CNSIM1","CNSIM.CNSIM2"))

Command to logout:

DSI::datashield.logout(connections)

The DataSHIELD approach: aggregate and assign functions

How assign and aggregate functions work

DataSHIELD commands call functions that range from carrying out pre-requisite tasks such as login to the datasources, to generating basic descriptive statistics, plots and tabulations. More advance functions allow for users to fit generalized linear models and generalized estimating equations models. R can list all functions available in DataSHIELD.

This section explains the functions we will call during this tutorial. Although this knowledge is not required to run DataSHIELD analyses it helps to understand the output of the commands. It can explain why some commands call functions that return nothing to the user, but rather store the output on the server of the data provider for use in a second function.

In DataSHIELD the person running an analysis (the client) uses client-side functions to issue commands (instructions). These commands initiate the execution (running) of server-side functions that run the analysis server-side (behind the firewall of the data provider). There are two types of server-side function: assign functions and aggregate functions.

Assign functions do not return an output to the client, with the exception of error or status messages. Assign functions create new objects and store them server-side either because the objects are potentially disclosive, or because they consist of the individual-level data which, in DataSHIELD, is never seen by the analyst. These new objects can include:

new transformed variables (e.g. mean centred or log transformed variables)
a new variable of a modified class (e.g. a variable of class numeric may be converted into a factor which R can then model as having discrete categorical levels)
a subset object (e.g. a dataframe including gender as a variable may be split into males and females).

Assign functions return no output to the client except to indicate an error or useful messages about the object store on server-side.

Aggregate functions analyse the data server-side and return an output in the form of aggregate data (summary statistics that are not disclosive) to the client. The help page for each function tells us what is returned and when not to expect an output on client-side.

Basic statistics and data manipulations

Descriptive statistics: variable dimensions and class

It is possible to get some descriptive or exploratory statistics about the assigned variables held in the server-side R session such as number of participants at each data provider, number of participants across all data providers and number of variables. Identifying parameters of the data will facilitate your analysis.

connections <- DSI::datashield.login(logins = logindata, assign = TRUE, symbol = "DST")
ds.dim(x = 'DST')

The output of the command is shown below. It shows that in study 1 there are 2163 individuals with 11 variables and in study 2 there are 3088 individuals with 11 variables, and that in both studies together there are in total 5251 individuals with 11 variables:

Aggregated (dimDS("DST")) [==============================================================] 100% / 0s
$`dimensions of DST in study1`
[1] 2163   11

$`dimensions of DST in study2`
[1] 3088   11

$`dimensions of DST in combined studies`
[1] 5251   11

Almost all functions in DataSHIELD can display split results (results separated for each study) or pooled results (results for all the studies combined). This can be done using the argument type='split' or type='combine' in each function. The majority of DataSHIELD functions default to type='combine'. The default for each function can be checked in the function help page. Some of the new versions of functions include the option type='both' which returns both the split and the pooled results.

Up to here, the dimensions of the assigned data frame DST have been found using the ds.dim command in which type='both' is the default argument.
Now use the type='combine' argument in the ds.dim function to identify the number of individuals (5251) and variables (11) pooled across all studies:

ds.dim(x='DST', type='combine', datasources = connections)

  Aggregated (dimDS("DST")) [==============================================================] 100% / 0s
$`dimensions of DST in combined studies`
[1] 5251   11

The argument "datasources=" is routinely specified in this tutorial for the purpose of clarity; however it can be omitted in general DataSHIELD practice- if the datasources argument is not specified the default set of connections will be used.

To check the variables in each study are identical (as is required for pooled data analysis), use the ds.colnames function on the assigned data frame D:

ds.colnames(x='DST', datasources = connections)

  Aggregated (exists("DST")) [=============================================================] 100% / 1s
  Aggregated (classDS("DST")) [============================================================] 100% / 1s
  Aggregated (colnamesDS("DST")) [=========================================================] 100% / 0s
$study1
 [1] "LAB_TSC"            "LAB_TRIG"           "LAB_HDL"            "LAB_GLUC_ADJUSTED"  "PM_BMI_CONTINUOUS" 
 [6] "DIS_CVA"            "MEDI_LPD"           "DIS_DIAB"           "DIS_AMI"            "GENDER"            
[11] "PM_BMI_CATEGORICAL"

$study2
 [1] "LAB_TSC"            "LAB_TRIG"           "LAB_HDL"            "LAB_GLUC_ADJUSTED"  "PM_BMI_CONTINUOUS" 
 [6] "DIS_CVA"            "MEDI_LPD"           "DIS_DIAB"           "DIS_AMI"            "GENDER"            
[11] "PM_BMI_CATEGORICAL"

Use the ds.class function to identify the class (type) of a variable - for example if it is an integer, character, factor etc. This will determine what analysis you can run using this variable class. The example below defines the class of the variable LAB_HDL held in the assigned data frame D, denoted by the argument x='DST$LAB_HDL'.

ds.class(x='DST$LAB_HDL', datasources = connections)

  Aggregated (exists("DST")) [=============================================================] 100% / 0s
  Aggregated (classDS("DST$LAB_HDL")) [====================================================] 100% / 1s
$study1
[1] "numeric"

$study2
[1] "numeric"

Descriptive statistics: quantiles and mean

As LAB_HDL is a numeric variable the distribution of the data can be explored.

The function ds.quantileMean returns the quantiles and the statistical mean.

It does not return minimum and maximum values as these values are potentially disclosive (e.g. the presence of an outlier). By default type='combined' in this function - the results reflect the quantiles and mean pooled for all studies. Specifying the argument type='split' will give the quantiles and mean for each study:

ds.quantileMean(x='DST$LAB_HDL', datasources = connections)

  Aggregated (exists("DST")) [=============================================================] 100% / 0s
  Aggregated (classDS("DST$LAB_HDL")) [====================================================] 100% / 1s
  Aggregated (quantileMeanDS(DST$LAB_HDL)) [===============================================] 100% / 0s
  Aggregated (lengthDS("DST$LAB_HDL")) [===================================================] 100% / 0s
  Aggregated (numNaDS(DST$LAB_HDL)) [======================================================] 100% / 0s
 Quantiles of the pooled data
       5%       10%       25%       50%       75%       90%       95%      Mean 
0.8606589 1.0385205 1.2964949 1.5704848 1.8418712 2.0824057 2.2191369 1.5619572

To get the statistical mean alone, use the function ds.mean use the argument type to request split results:

ds.mean(x='DST$LAB_HDL', datasources = connections)

  Aggregated (meanDS(DST$LAB_HDL)) [=======================================================] 100% / 0s
$Mean.by.Study
       EstimatedMean Nmissing Nvalid Ntotal
study1      1.569416      360   1803   2163
study2      1.556648      555   2533   3088

$Nstudies
[1] 2

$ValidityMessage
       ValidityMessage 
study1 "VALID ANALYSIS"
study2 "VALID ANALYSIS"

Conclusion