Hi, Peng. I'm including the bioc list so that everyone benefits from the answer. I hope you don't mind. See answers below. On Sat, Apr 30, 2011 at 6:38 PM, Peng Yu <pengyu.ut@gmail.com> wrote: > Hi Sean, > > Some matrix files are very big. > > > ftp://ftp.ncbi.nih.gov/pub/geo/DATA/SeriesMatrix/GSE18927/GSE18927-G PL5188_series_matrix.txt.gz > > I don't need anything after 91 in that file. I create a new file that > only have the first 90 lines. > 91 !series_matrix_table_begin > > But getGEO gives me the following error. > > > gse=getGEO(file='GSE18927-GPL5188_series_matrix_reduced.txt.gz') > Error in read.table(file = file, header = header, sep = sep, quote = quote, > : > no lines available in input > > Unfortunately, the GSE series file has a specific format, so just removing parts of it is likely to break parsing as your example shows. You could, of course, just read your edited file into R directly if that is the route you want to go. > Usually, I only need the meta data for samples but not the data matrix > like the one in the above example. Is there a way to exclude the > unwanted information when parsing the file? Take a look at the GEOmetadb package. That is the fastest way to get the metadata from GEO that I know of. It contains nearly ALL the metadata in GEO parsed into a SQLite database that is updated about weekly. Sean [[alternative HTML version deleted]]