88485cfa7f13affa28ec5765fe9b2db90cce42b6 markd Tue Dec 8 21:02:15 2020 -0800 hgPcr working diff --git src/webBlat/install.txt src/webBlat/install.txt index 69b13fb..95d7565 100644 --- src/webBlat/install.txt +++ src/webBlat/install.txt @@ -55,33 +55,33 @@ ssh bigRamMachine cd /data/genomes/twoBit gfServer start bigRamMachine 17779 hg19.2bit & gfServer -trans -mask start bigRamMachine 17778 hg19.2bit & the -trans flag makes a translated index. It will take approximately 15 minutes to build an untranslated index, and 45 minutes to build a translate index. To build an untranslated index to be shared with In-silico PCR do gfServer -stepSize=5 bigRamMachine 17779 hg19.2bit & This index will be slightly more sensitive, noticeably so for small query sequences, with blat. CREATING A DYNAMIC GFSERVER -To avoid having a larger number of gfServer processing sitting in memory, -indices can be pre-generate as files on disk and then gfServer instances -launched dynamically using xinetd. +To avoid having a number of gfServer processing sitting in memory, indices can +be pre-generate as files on disk and then gfServer instances launched +dynamically using xinetd. See gfServer/blat-xinetd.example for an example xinetd configuration file. Data files must be create using the following layout, where the $rootdir is the directory specified to gfServer in the xinetd file $rootdir/$genome/$genome.2bit $rootdir/$genome/$genome.untrans.gfidx $rootdir/$genome/$genome.trans.gfidx The gfServer index command is used to generate the indices: gfServer index $rootdir/$genome/$genome.untrans.gfidx $rootdir/$genome/$genome.2bit gfServer index -trans $rootdir/$genome/$genome.trans.gfidx $rootdir/$genome/$genome.2bit