Update LDSC output format

Al-Murphy · Jan 15, 2024 · cccf77b · cccf77b
1 parent c55aaa7
commit cccf77b
Show file tree

Hide file tree

Showing 11 changed files with 98 additions and 20 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7 +1,7 @@
 Package: MungeSumstats
 Type: Package
 Title: Standardise summary statistics from GWAS
-Version: 1.11.2
+Version: 1.11.3
 Authors@R:
     c(person(given = "Alan",
            family = "Murphy",

diff --git a/NEWS.md b/NEWS.md
@@ -1,3 +1,11 @@
+## CHANGES IN VERSION 1.11.3
+
+### Bug fix
+* For LDSC format, rename A1 and A2 as LDSC expects A1 to be the effect column 
+rather than A2 (the opposite to MSS's default) - see more [here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68).
+Although, this didn't seem to make any difference to results in tests, see more
+[here](https://github.com/neurogenomics/MungeSumstats/issues/160#issuecomment-1891899253).
+
 ## CHANGES IN VERSION 1.11.2
 
 ### Bug fix

diff --git a/R/format_sumstats.R b/R/format_sumstats.R
@@ -187,7 +187,12 @@
 #' @param ldsc_format DEPRECATED, do not use. Use save_format="LDSC" instead.
 #' @param save_format Output format of sumstats. Options are NULL - standardised
 #' output format from MungeSumstats, LDSC - output format compatible with LDSC
-#' and openGWAS - output compatible with openGWAS VCFs. Default is NULL.
+#' and openGWAS - output compatible with openGWAS VCFs. Default is NULL. 
+#' **NOTE** - If LDSC format is used, the naming convention of A1 as the 
+#' reference (genome build) allele and A2 as the effect allele will be reversed
+#' to match LDSC (A1 will now be the effect allele). See more info on this 
+#' [here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68). Note that any 
+#' effect columns (e.g. Z) will be inrelation to A1 now instead of A2.
 #' @param log_folder_ind Binary Should log files be stored containing all
 #' filtered out SNPs (separate file per filter). The data is outputted in the
 #' same format specified for the resulting sumstats file. The only exception to
@@ -285,8 +290,7 @@ format_sumstats <- function(path,
     #### Setup multi-threading ####
     data.table::setDTthreads(threads = nThread)
     #### Setup empty variables ####
-    rsids <- NULL
-    orig_dims <- NULL
+    rsids <- orig_dims <- A1_n <- A2 <- A1 <- NULL
     log_files <- vector(mode = "list")
     t1 <- Sys.time()
 
@@ -1036,6 +1040,21 @@ format_sumstats <- function(path,
         ### Check 39: Ensure CHR follows the requested style ###
         CHR <- NULL
         sumstats_return$sumstats_dt[, CHR := GenomeInfoDb::mapSeqlevels(CHR, style = chr_style)]
+
+        ### IF LDSC, rename A1 and A2, effect columns are fine
+        if (!is.null(save_format) && 
+            tolower(save_format)=="ldsc") {
+          message("Renaming A1,A2 to match LDSC format.")
+          #For LDSC format, rename A1 and A2 as LDSC expects A1 to be the effect 
+          #column rather than A2 (the opposite to MSS's default) - see more 
+          #[here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68).Although, 
+          #this didn't seem to make any difference to results in tests, see more
+          #https://github.com/neurogenomics/MungeSumstats/issues/160#issuecomment-1891899253
+          sumstats_return$sumstats_dt[,A1_n:=A2]
+          sumstats_return$sumstats_dt[,A2:=A1]
+          sumstats_return$sumstats_dt[,A1:=A1_n]
+          sumstats_return$sumstats_dt[,A1_n:=NULL]
+        }
 
         #### WRITE data.table TO PATH ####
         check_save_out$save_path <- write_sumstats(

diff --git a/README.md b/README.md
@@ -1,19 +1,22 @@
 `MungeSumstats`: Standardise the format of GWAS summary statistics
 ================
-<h5> ¶ <i>Authors</i>: Alan Murphy, Brian Schilder and Nathan Skene ¶
+<h5>  
+<i>Authors</i>: Alan Murphy, Brian Schilder and Nathan Skene  
+</h5>
+<h5>  
+<i>Updated</i>: Jan-15-2024  
 </h5>
-<h5> ¶ <i>Updated</i>: Jul-13-2023 ¶ </h5>
 
 <!-- Readme.md is generated from Readme.Rmd. Please edit that file -->
 <!-- badges: start -->
 
-[![](https://img.shields.io/badge/release%20version-1.8.0-black.svg)](https://www.bioconductor.org/packages/MungeSumstats)
-[![](https://img.shields.io/badge/devel%20version-1.9.11-black.svg)](https://github.com/neurogenomics/MungeSumstats)
+[![](https://img.shields.io/badge/release%20version-1.10.1-black.svg)](https://www.bioconductor.org/packages/MungeSumstats)
+[![](https://img.shields.io/badge/devel%20version-1.11.3-black.svg)](https://github.com/neurogenomics/MungeSumstats)
 [![R build
 status](https://github.com/neurogenomics/MungeSumstats/workflows/rworkflows/badge.svg)](https://github.com/neurogenomics/MungeSumstats/actions)
 [![](https://img.shields.io/github/last-commit/neurogenomics/MungeSumstats.svg)](https://github.com/neurogenomics/MungeSumstats/commits/master)
 [![](https://codecov.io/gh/neurogenomics/MungeSumstats/branch/master/graph/badge.svg)](https://codecov.io/gh/neurogenomics/MungeSumstats)
-[![](https://img.shields.io/badge/download-5460/total-blue.svg)](https://bioconductor.org/packages/stats/bioc/MungeSumstats)
+[![](https://img.shields.io/badge/download-11379/total-blue.svg)](https://bioconductor.org/packages/stats/bioc/MungeSumstats)
 [![License:
 Artistic-2.0](https://img.shields.io/badge/license-Artistic--2.0-blue.svg)](https://cran.r-project.org/web/licenses/Artistic-2.0)
 [![](https://img.shields.io/badge/doi-https://doi.org/10.1093/bioinformatics/btab665-blue.svg)](https://doi.org/https://doi.org/10.1093/bioinformatics/btab665)
@@ -154,10 +157,10 @@ We would like to acknowledge all those who have contributed to
 
 <div id="ref-Skene2018" class="csl-entry">
 
-<span class="csl-left-margin">1. </span><span
-class="csl-right-inline">Nathan G. Skene, T. E. B., Julien Bryois.
-Genetic identification of brain cell types underlying schizophrenia.
-*Nature Genetics* (2018).
+<span class="csl-left-margin">1.
+</span><span class="csl-right-inline">Nathan G. Skene, T. E. B., Julien
+Bryois. Genetic identification of brain cell types underlying
+schizophrenia. *Nature Genetics* (2018).
 doi:[10.1038/s41588-018-0129-5](https://doi.org/10.1038/s41588-018-0129-5)</span>
 
 </div>

diff --git a/man/check_ldsc_format.Rd b/man/check_ldsc_format.Rd
diff --git a/man/format_sumstats.Rd b/man/format_sumstats.Rd
diff --git a/man/import_sumstats.Rd b/man/import_sumstats.Rd
diff --git a/man/validate_parameters.Rd b/man/validate_parameters.Rd
diff --git a/man/write_sumstats.Rd b/man/write_sumstats.Rd
diff --git a/tests/testthat/test-vcf_formatting.R b/tests/testthat/test-vcf_formatting.R
@@ -146,12 +146,26 @@ test_that("VCF is correctly formatted", {
         ldsc_cols <- c("SNP", "N", "A1", "A2", "Z")
         testthat::expect_true(all(ldsc_cols %in% names(res)))
 
+        #also ensure A1 and A2 have been renamed
+        #For LDSC format, rename A1 and A2 as LDSC expects A1 to be the effect 
+        #column rather than A2 (the opposite to MSS's default) - see more 
+        #[here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68).Although, 
+        #this didn't seem to make any difference to results in tests, see more
+        #[here](https://github.com/neurogenomics/MungeSumstats/issues/160#issuecomment-1891899253).
+        data.table::setnames(res,c("A1","A2"),c("A2","A1"))
+        res[,CHR:=as.character(CHR)]
+        testthat::expect_true(all.equal(res[,c("SNP","CHR","BP","A1","A2","END",
+                                               "FILTER","FRQ","BETA","LP","SE",
+                                               "P")],rtrn_dt))
+
+
         testthat::expect_equal(reformatted_lines[1:5], corr_res)
     } else {
         testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
         testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
         testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
         testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
         testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
+        testthat::expect_true((is_32bit_windows||!Sys.info()["sysname"]=="Linux"))
     }
 })
diff --git a/vignettes/MungeSumstats.Rmd b/vignettes/MungeSumstats.Rmd
@@ -165,7 +165,11 @@ flexibility to export the reformatted file as tab-delimited, VCF or R
 native objects such as data.table, GRanges or VRanges objects. The
 output can also be outputted in an **LDSC ready** format which means the
 file can be fed directly into LDSC without the need for additional
-munging.
+munging. **NOTE** - If LDSC format is used, the naming convention of A1 as the 
+reference (genome build) allele and A2 as the effect allele will be reversed
+to match LDSC (A1 will now be the effect allele). See more info on this 
+[here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68). Note that any 
+effect columns (e.g. Z) will be inrelation to A1 now instead of A2.
 
 # Data
 
@@ -419,7 +423,12 @@ conducted by *MungeSumstats* are:
     ("data.table","vranges","granges").
 -   **save_format** Ensure that output format meets all requirements to
     be passed directly into LDSC ("ldsc") without the need for additional
-    munging or for IEU OpenGWAS format ("opengwas") before saving as a VCF
+    munging or for IEU OpenGWAS format ("opengwas") before saving as a VCF.
+    **NOTE** - If LDSC format is used, the naming convention of A1 as the 
+    reference (genome build) allele and A2 as the effect allele will be reversed
+    to match LDSC (A1 will now be the effect allele). See more info on this 
+    [here](https://groups.google.com/g/ldsc_users/c/S7FZK743w68). Note that any 
+    effect columns (e.g. Z) will be inrelation to A1 now instead of A2.
 -   **log_folder_ind** Should log files be stored containing all
     filtered out SNPs (separate file per filter). The data is outputted
     in the same format specified for the resulting sumstats file.