Trim (#35)

* file out

* cleaning

README.Rmd

@@ -139,6 +139,22 @@ sets_10x_withmeta <- sets_10x %>%
     str_detect(str_to_lower(samplemetainfo_1), "meta|annot|type|clustering|coldata") | 
     str_detect(str_to_lower(additionalfiles), "meta|annot|type|clustering|coldata")
   )
+
+sets_final <- sets_10x %>% 
+  mutate(has_meta = ifelse(
+    str_detect(str_to_lower(samplemetainfo_1), "meta|annot|type|clustering|coldata") | 
+    str_detect(str_to_lower(additionalfiles), "meta|annot|type|clustering|coldata"),
+    "yes",
+    "no")) %>% 
+  mutate(has_meta = ifelse(
+    is.na(has_meta), "no", has_meta
+  )) %>% 
+  select(ID, ReleaseDate, PubmedID, Species, 
+         files = samplemetainfo_1, type = samplemetainfo_2,
+         has_meta) %>% 
+  filter(str_sub(ReleaseDate, 1, 4) != "2021")
+
+# write_tsv(sets_final, "arrayexpres_analysis_2020.tsv.gz")
 ```
 
 Current fraction in GEO with metadata: **`r (gds$usable == "yes") %>% mean(na.rm = TRUE)`**. In comparison, for ArrayExpress 10x datasets, the fraction is **`r nrow(sets_10x_withmeta) / nrow(sets_10x)`**).

inst/manuscript/data_readme.txt

@@ -0,0 +1,37 @@
+Numeric data underlying Figure 2 and analyses of Puntambekar et al. 2021
+
+fig2a.tsv:
+id : GEO accession
+year : year of publication
+usable : single cell dataset has metadata
+usable2 : format the metadata stored in
+
+fig2b.tsv:
+id : GEO accession
+journal	: published journal
+journal_group : subgrouping of journal
+
+fig2c.tsv:
+id : GEO accession
+usable : single cell dataset has metadata
+journal	: published journal
+cite : number of citations
+year : year of publication
+ifs : journal impact factor
+
+fig2d.tsv:
+id : GEO accession
+usable : single cell dataset has metadata
+usable2 : format the metadata stored in
+software_author : whether an author is involved with scRNA-seq software development
+
+geo_vs_Svensson.tsv:
+id : GEO accession
+useable : single cell dataset has metadata. Entries in Svensson but not found in automated search are labeled NA
+
+manual_check.tsv:
+id : GEO accession
+is_sc : manual check if entry is truly single cell data
+called_usable : orginal automated call if single cell dataset has metadata
+correct_call : manual check if automated call was correct
+has_type : manual check if dataset metadata actually contains cell type column/info

utils/geo_query.R

@@ -90,7 +90,6 @@ get_date <- function(res) {
 }
 
 get_pubmed <- function(res) {
-  # res <- res[[1]]
   if (class(res) != "list") {
     return(NA)
   }
@@ -116,12 +115,6 @@ get_pubmed <- function(res) {
     return(NA)
   }
   return(id2)
-  # tryCatch(
-  #   get_pubmed_ids(id2) %>%
-  #     fetch_pubmed_data() %>%
-  #     article_to_df(),
-  #   error = function(e) {"error"}
-  # )
 }
 
 from_list <- function(pubmedid) {