fix: expand() handles NULL geno data in sites-only VCFs (#72)

DESCRIPTION

@@ -3,7 +3,7 @@ Type: Package
 Title: Annotation of Genetic Variants
 Description: Annotate variants, compute amino acid coding changes,
         predict coding outcomes.
-Version: 1.59.0
+Version: 1.59.1
 Authors@R: c(
         person("Valerie", "Oberchain", role="aut"),
         person("Martin", "Morgan", role="aut"),

NEWS

@@ -1,3 +1,16 @@
+CHANGES IN VERSION 1.59.1
+-------------------------
+
+BUG FIXES
+
+    o expand() no longer errors with "'data' must be of a vector type,
+      was 'NULL'" when called on sites-only VCFs (e.g., gnomAD allele
+      frequency files) that declare FORMAT fields in the header but have
+      no sample columns. The internal .expandAD() helper now returns an
+      appropriately-shaped empty array when the geno data is an empty
+      list (0 samples), instead of passing NULL to array(). (GitHub
+      issue #72)
+
 CHANGES IN VERSION 1.36.0
 -------------------------
 

R/methods-expand.R

@@ -118,6 +118,12 @@ setMethod("expand", "CollapsedVCF",
 .expandAD <- function(AD, idxlen, xcols)
 {
     if (is.list(AD)) {
+        ## No data to expand (e.g., sites-only VCF with 0 samples;
+        ## list elements may all be NULL). Return empty array.
+        ## (GitHub issue #72)
+        if (length(AD) == 0L || xcols == 0L)
+            return(array(integer(0L), c(idxlen, xcols, 2L)))
+
         adpart <- PartitioningByWidth(AD)
         if (any(zeros <- width(adpart) == 0L)) { 
             AD[zeros] <- list(rep(NA_integer_, 2L))

inst/unitTests/cases/sites_only.vcf

@@ -0,0 +1,9 @@
+##fileformat=VCFv4.2
+##INFO=<ID=AF,Number=A,Type=Float,Description="Allele Frequency">
+##INFO=<ID=DP,Number=1,Type=Integer,Description="Total Depth">
+##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
+##FORMAT=<ID=AD,Number=R,Type=Integer,Description="Allelic depths">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
+chr1	100	rs1	A	C,T	50	PASS	AF=0.1,0.2;DP=100
+chr1	200	rs2	G	A	30	PASS	AF=0.3;DP=50
+chr1	300	rs3	T	G,C,A	99	PASS	AF=0.4,0.3,0.1;DP=200

inst/unitTests/test_expand-methods.R

@@ -90,3 +90,34 @@ test_expand_adr_adf <- function()
 			})
 
 }
+
+## ---------------------------------------------------------------
+## Test for GitHub issue #72: expand() errors with 'data' must be
+## of a vector type, was 'NULL' on sites-only VCFs (e.g. gnomAD).
+## ---------------------------------------------------------------
+test_expand_sitesOnly_issue72 <- function()
+{
+    fl <- system.file("unitTests", "cases", "sites_only.vcf",
+                      package="VariantAnnotation")
+    vcf <- readVcf(fl, "")
+
+    ## Verify this is indeed a sites-only VCF (0 samples)
+    checkIdentical(ncol(vcf), 0L)
+    checkTrue(nrow(vcf) > 0L)
+
+    ## expand() must not error on sites-only VCF with FORMAT headers
+    exp <- expand(vcf)
+    checkTrue(is(exp, "ExpandedVCF"))
+
+    ## Multi-allelic rows are expanded correctly
+    ## Input: 3 rows with ALT counts 2, 1, 3 => 6 expanded rows
+    checkIdentical(nrow(exp), 6L)
+    checkIdentical(ncol(exp), 0L)
+
+    ## INFO Number=A fields are expanded and scalar
+    checkTrue(is.numeric(info(exp)$AF))
+    checkEquals(info(exp)$AF, c(0.1, 0.2, 0.3, 0.4, 0.3, 0.1))
+
+    ## Geno data is preserved (empty but valid structure)
+    checkTrue(length(geno(exp)) >= 0L)
+}