scripts/DA.R

[jenna-a2ai]

Data Assembly

Note: Please modify the checklist items to insert relevant QC context.

Reproducibility and Organization

• Script is a quarto document.
• Script renders free of error within the Rproject space.
• renv::status() present at beginning of script and passes.
• Relative paths are used.
• .Rprofile is present in same directory as final script.
• 🔴 Parquet files are used where available.
  • files that should be parquet: [INSERT]
• 🔴 Data hash codes are printed and match source and output data.
  • expected hash files: [INSERT]
• 🔴 Correct source data files are read in.
  • correct source data: [INSERT]
• Final dataset outputs to expected location as a csv, converts all NA values to a period, and outputs a hash.
• All code chunks are labeled.
• session.info() included at the conclusion of script.

yspec

• All yspec files can be found in <project>/data/derived and are excluded from gitignore.
• All variables have descriptions. All descriptions are easy to read and understand. All variables have units where applicable.
• All categorical variables have all options outlined and a 1:1 numerical decoded value. All categorical variables match the decoding present in the yspec.
• 🔴 All continuous variables specify units and all units match those of the final dataset. If multiple analytes are included in a single column, units are clear for each analyte.
• 🔴 The yspec file contains all necessary variables for analysis, including both project specific variables and NONMEM required variables.

Dataset Assumptions - Automated

🔴 These can be done using pointblank and QCer can just confirm these pass!

• 🔴 LIST POINTBLANK ASSUMPTIONS.
• The final dataset includes both a numeric (ID) and character (e.g. USUBJID) version of the unique subject identifier.
• The final dataset includes all columns listed in the corresponding yspec file.
• Any missing columns or observations are confirmed to be missing in source data and marked as -999.
• Missing data is documented appropriately in dataset.

Data Assumptions - Manual

🔴 These CAN’T be done using pointblank and QCer must look closely at specific sections of code.

• 🔴 All manual conversions are accurate and consistent.
• 🔴 All manual imputations are well documented to outside sources or sufficiently justified.
• 🔴 All manual imputations are accurate according to their specified source.
• 🔴 All assumptions are well justified (e.g. by the protocol, project lead decision, client decision, etc.), documented, and reasonable.
• 🔴 All records to be excluded (e.g. for all analyses, for sensitivity analysis only) are properly identified, justified, and documented (i.e., CFLAG or other flag column).

Misc

• Code is easily readable and follows general coding guidelines (link to code hub).
• Helper functions are well documented, make sense, make reasonable assumptions, and follow good coding practices.
• 🔴 Helper function arguments match as stated in yspec (e.g. SEX variable in Cockcroft-Gault equation matches decoding in yspec).

Metadata

• author: jenna-a2ai [email protected]
• qc type: Data Assembly
• script hash: 9de0992cc46272ab53f38ac5f79fe6f4
• git sha: aced09b
• file history: https://github.com/A2-ai/gilead_ghqc_demo/commits/main/scripts/DA.R

[wes-a2ai]

gilead_ghqc_demo/scripts/DA.R

Line 19 in aced09b

select(C, LINE, ID, ATFD, ATLD, NTFD, NTLD, TEST, AMT, DV, LDV, EVID, CMT, MDV, BLQ, LLOQ, DOSEA, IMPDOS, STUDY, USUBJID, DTIM, FDOSE, VISIT, TPT, CFLAG)

Chose the wrong time column

[jenna-a2ai]

@wes-a2ai

thanks, Wes! Nominal time, right?

File Difference

reference commit (older version): aced09b

comparator commit (newer version): 3bacb81

@@ - reference:  lines 17-21 @@
@@ + comparator: lines 17-21 @@
  17     CFLAG = 100                                                                                                                                            
  18   ) %>%                                                                                                                                                    
- 19   select(C, LINE, ID, ATFD, ATLD, NTFD, NTLD, TEST, AMT, DV, LDV, EVID, CMT, MDV, BLQ, LLOQ, DOSEA, IMPDOS, STUDY, USUBJID, DTIM, FDOSE, VISIT, TPT, CFLAG)
+ 19   select(C, LINE, ID, NTFD, NTLD, NTFD, NTLD, TEST, AMT, DV, LDV, EVID, CMT, MDV, BLQ, LLOQ, DOSEA, IMPDOS, STUDY, USUBJID, DTIM, FDOSE, VISIT, TPT, CFLAG)
  20 write.csv(pkdf5, file.path(derDir, "PK_AXAN101.csv"), na = ".")                                                                                            
  21                                                                                                                                                            

Metadata

• reference commit: aced09b
• reference script hash: 5eb233c1590493e6d7b3e550e563f70f
• comparator commit: 3bacb81
• comparator script hash: 05c92a0662bfbcddd38d1ef3404f8715