Simple bash script to download publication tables from ucsc genome browser and merge.

setup_ucsc_pub_files.sh

# Downloading pub tables. Requires wget. Can be installed using home brew for mac
# More information is available here: http://brew.sh/

mkdir ../data/

# Download Files
wget  --timestamping --directory-prefix='../data/' 'http://hgdownload.cse.ucsc.edu/goldenPath/hgFixed/database/pubsArticle.txt.gz'
wget  --timestamping --directory-prefix='../data/' 'http://hgdownload.cse.ucsc.edu/goldenPath/hgFixed/database/pubsMarkerAnnot.txt.gz'
## wget  --timestamping --directory-prefix='../data/' 'http://hgdownload.cse.ucsc.edu/goldenPath/hgFixed/database/pubsSequenceAnnot.txt.gz'

# Unzip Files
gunzip ../data/pubsArticle.txt.gz
gunzip ../data/pubsMarkerAnnot.txt.gz
## gunzip ../data/pubsSequenceAnnot.txt.gz

# Cut out needed columns from files. Sort, and make unique.
cut -f 1,5,6,8  pubsMarkerAnnot.txt | sort -n -k 1 | uniq -u > pubsMarkerAnnot_cut.txt
cut -f 1,2,3,8  pubsArticle.txt | sort -n -k 1 | uniq -u > pubsArticle_cut.txt

# Join the cut files, rearrange columns and strip the <B> and </B> Tags. Also removes the article index.
# Apparently, the tab needs to be specified as a literal (hence the $ sign).
join -t $'\t'  pubsMarkerAnnot_cut.txt pubsArticle_cut.txt | uniq -u | awk -F $'\t' '{print $6"\t"$5"\t"$2"\t"$3"\t"$7"\t"$4}' | sed "s/<B>//;s/<\/B>//" > pubs_join.txt

# columns:

# pmid
# pmc id
# marker type (gene, snp, band)
# marker name (e.g. BRCA1)
# publication title 
# Snippet

# Create the 'unique titles', and 'unique snippets' files:
cut -f 5 pubs_join.txt | uniq > titles_unique.txt
cut -f 6 pubs_join.txt | uniq > snippets_unique.txt

Thanks Dan!