Merge pull request YuLab-SMU#51 from huerqiang/master

upate data
huerqiang · Aug 16, 2021 · 0a9daef · 0a9daef
2 parents 2d055a4 + 78ade9a
commit 0a9daef
Show file tree

Hide file tree

Showing 13 changed files with 14 additions and 3 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7 +1,7 @@
 Package: DOSE
 Type: Package
 Title: Disease Ontology Semantic and Enrichment analysis
-Version: 3.19.2
+Version: 3.19.2.991
 Authors@R: c( person(given = "Guangchuang", family = "Yu",        email = "[email protected]",   role = c("aut", "cre")),
               person(given = "Li-Gen",      family = "Wang",      email = "[email protected]",   role = "ctb"),
               person(given = "Vladislav",   family = "Petyuk",    email = "[email protected]",          role = "ctb"),

diff --git a/NEWS.md b/NEWS.md
@@ -1,3 +1,7 @@
+# DOSE 3.19.2.991
+
++ upate DisGeNET and NCG data(2021-8-16, Mon)
+
 # DOSE 3.19.2
 
 + bug fixed, change 'is.na(path2name)' to 'all(is.na(path2name))' (2021-06-21, Mon)

diff --git a/data/DGN_EXTID2PATHID.rda b/data/DGN_EXTID2PATHID.rda
diff --git a/data/DGN_PATHID2EXTID.rda b/data/DGN_PATHID2EXTID.rda
diff --git a/data/DGN_PATHID2NAME.rda b/data/DGN_PATHID2NAME.rda
diff --git a/data/NCG_EXTID2PATHID.rda b/data/NCG_EXTID2PATHID.rda
diff --git a/data/NCG_PATHID2EXTID.rda b/data/NCG_PATHID2EXTID.rda
diff --git a/data/NCG_PATHID2NAME.rda b/data/NCG_PATHID2NAME.rda
diff --git a/data/VDGN_EXTID2PATHID.rda b/data/VDGN_EXTID2PATHID.rda
diff --git a/data/VDGN_PATHID2EXTID.rda b/data/VDGN_PATHID2EXTID.rda
diff --git a/data/VDGN_PATHID2NAME.rda b/data/VDGN_PATHID2NAME.rda
diff --git a/inst/extdata/build_DGN_Anno.R b/inst/extdata/build_DGN_Anno.R
@@ -1,4 +1,8 @@
-x <- read.delim("all_gene_disease_associations.tsv", comment.char="#", stringsAsFactor=F)
+# download from https://www.disgenet.org/downloads
+# 1. ALL gene-disease associations
+# 2. ALL variant-disease associations
+#x <- read.delim("all_gene_disease_associations.tsv", comment.char="#", stringsAsFactor=F)
+x <- read.delim("all_gene_disease_associations.tsv", comment.char="#", stringsAsFactor=F, encoding = "latin1")
 d2n <- unique(x[, c("diseaseId", "diseaseName")])
 d2g <- unique(x[, c("diseaseId", "geneId")])
 

diff --git a/inst/extdata/build_NCG_Anno.R b/inst/extdata/build_NCG_Anno.R
@@ -1,4 +1,7 @@
-x=read.delim("NCG6_cancergenes.tsv", stringsAsFactor=F)
+# download from http://ncg.kcl.ac.uk/download.php
+# All cancer genes -> •List of 2372 cancer genes and supporting literature
+#x=read.delim("NCG6_cancergenes.tsv", stringsAsFactor=F)
+x=read.delim("NCG6_cancergenes.tsv", stringsAsFactor=F, encoding = "latin1")
 path2gene <- x[, c("cancer_type", "entrez")]
 path2gene <- path2gene[path2gene[,1] != '',]