[BioC] paper - download - pubmed

[Chris Stubben]

Sorry, not sure why references[1] were added automatically to html links 
within the code, but this reply should work if you copy and paste (I hope).
Chris

 >>
 >> So, the problem is not that, for each paper I have to download the
 >> pdfs (which are available if I go to the pubmed and search directly
 >> there) and the corresponding supplementary files.
 >>
Nooshin,
You can download pdfs from Pubmed Central if you have one PMC id.
download.file( 
"http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3446303/pdf", "PMC3446303.pdf")

However, NCBI clearly states that you may NOT use any kind of automated 
process to download articles in bulk from the main PMC site, so I would 
use the ftp site for Open Access articles (see 
http://www.ncbi.nlm.nih.gov/pmc/tools/ftp ).  The ftp site also has the 
supplemental files included.  First, read the list of available files

pmcftp <- read.delim( "ftp://ftp.ncbi.nlm.nih.gov/pub/pmc/file_list.txt" 
, skip=1, header=FALSE, stringsAsFactors=FALSE)
nrow(pmcftp)
[1] 552677
names(pmcftp)<-c("dir", "citation", "id")

Then match PMC ids and loop through the results to download and untar 
the files

y <- subset(pmcftp, id %in% c("PMC3446303", "PMC3463124") )
y
509377          75/e9/Genome_Biol_2012_Apr_24_13(4)_R29.tar.gz          
Genome Biol. 2012 Apr 24; 13(4):R29 PMC3446303
514389 04/0f/Bioinformatics_2012_Oct_1_28(19)_2532-2533.tar.gz 
Bioinformatics. 2012 Oct 1; 28(19):2532-2533 PMC3463124

for( i in 1: nrow(y) ){
destfile <- paste(y$id[i], ".tar.gz", sep="")
download.file( paste("ftp://ftp.ncbi.nlm.nih.gov/pub/pmc", y$dir[i], 
sep="/"),  destfile )
untar( destfile, compressed=TRUE)
}

Also, if you need to get a list of PMC ids in R,  I have a package 
called genomes on BioC that includes E-utility scripts. So something 
like this query would get the 49 pmc ids for articles with Bioconductor 
in the title.

x2 <- esummary(esearch("bioconductor[TITLE] AND open access[FILTER]", 
db="pmc"), version="2.0")

Esummary uses a generic parser by default, so PMCids are mashed together 
in a column with other Ids
ids <-gsub(".*(PMC[0-9]*)", "\\1", x2$ArticleIds)
y <- subset(pmcftp, id %in% ids)

You could run esummary and add parse=FALSE to get the XML results and 
parse that any way you like.  Or even use esearch and set usehistory="n"
ids2 <- paste("PMC",  esearch("bioconductor[TITLE] AND open 
access[FILTER]", db="pmc", usehistory="n", retmax=100), sep="")


-- 

Chris Stubben

Los Alamos National Lab
Bioscience Division
MS M888
Los Alamos, NM 87545