Running GGIR on the UK Biobank Research Analysis Platform

[Gavaguy]

Hello everyone,

I am trying to run the GGIR command on the UKB-Rep using Jupyter Lab.
I am trying to calculate the sleep regularity index.

I am using those accelerometer files, all in .csv format

file_list <- system("dx ls Bulk/Activity/Epoch/10/", intern = TRUE)

I am using the command

For externally derived UK Biobank data in .CSV format:

   GGIR(datadir = "/opt/notebooks",
       outputdir = "/opt/myoutput",
       dataFormat = "ukbiobank_csv",
       extEpochData_timeformat = "%m/%d/%Y %H:%M:%S",
       mode = c(1:2),
       do.report = c(2),
       windowsizes = c(5, 900, 3600), # We know that data was stored in 5 second epoch
       desiredtz = "Europe/London", # We know that data was collected in the UK
       visualreport = FALSE,
       overwrite = TRUE)

I seems that the part 1 output is generated, the output is

Checking that user has write access permission for directory specified by argument outputdir: Yes

GGIR version: 3.1.5

Do not forget to cite GGIR in your publications via a version number and
Migueles et al. 2019 JMPB. doi: 10.1123/jmpb.2018-0063.
See also: https://cran.r-project.org/package=GGIR/vignettes/GGIR.html#citing-ggir

To make your research reproducible and interpretable always report:
(1) Accelerometer brand and product name
(2) How you configured the accelerometer
(3) Study protocol and wear instructions given to the participants
(4) GGIR version
(5) How GGIR was used: Share the config.csv file or your R script.
(6) How you post-processed / cleaned GGIR output
(7) How reported outcomes relate to the specific variable names in GGIR

---

Part 1

P1 file: 1 2 3 4 5 6 7 8 9 10 11 12 …

However, then an error message occurs:

Error in seq.default(quartlystart, quartlystart + ((nrow(D) - 1) * epSizeShort), : 'from' must be a finite number
Traceback:

1. GGIR(datadir = "/opt/notebooks", outputdir = "/opt/myoutput",
   . dataFormat = "ukbiobank_csv", extEpochData_timeformat = "%m/%d/%Y %H:%M:%S",
   . mode = c(1:2), do.report = c(2), windowsizes = c(5, 900,
   . 3600), desiredtz = "Europe/London", visualreport = FALSE,
   . overwrite = TRUE)
2. convertEpochData(datadir = datadir, metadatadir = metadatadir,
   . params_general = params_general, verbose = verbose)
3. seq(quartlystart, quartlystart + ((nrow(D) - 1) * epSizeShort),
   . by = epSizeShort)
4. seq.default(quartlystart, quartlystart + ((nrow(D) - 1) * epSizeShort),
   . by = epSizeShort)
5. stop("'from' must be a finite number")

Do you have an idea what the issue is?

Do you perhaps have a complete code how the calculate the SRI in the UKB-Rep from beginning to end?

Thank you for your help in advance.

Kind regards

[vincentvanhees]

@Gavaguy

For externally derived UK Biobank data in .CSV format:

Please confirm that you are working with the UK Biobank csv file in 5 second resolution, because that is what this functionality is intended for.

If yes, then with this data you will not be able to derive SRI because it only includes ENMO values and no other signal features, which makes it impossible to do anything related to sleep.

[vincentvanhees]

Action point for me:

• Emphasize with big capitals in https://wadpac.github.io/GGIR/articles/Cookbook.html that those examples are specific to EXTERNALLY derived data, meaning epoch aggregates by external software or hardware. All of it does NOT relate to processing RAW accelerometer data.