Script to parse fasta headers

http://www.biostars.org/p/62884/

Assuming that you do not have spaces in your sequences you can try that:

sed -e '/^>/ s/ .*//' mybigfile

Awk should work too (even if your sequences contain spaces):

awk '{print /^>/ ? $1 : $0}' mybigfile

and just for fun, a pure bash version:

while read l ; do echo "${l%% *}" ; done < mybigfile