Good Evening: With these large bed files, you are going to have trouble with any of the formats unless you have a very large memory machine. Large as in 250 Gb of memory or more.
I would suggest breaking up your bed file per chromosome. Load the parts with table names: chr1_trackName chr22_trackName ... etc ... These tracks will function as a single track by the name of "trackName" Use "trackName" in your trackDb entry. This is called a 'split' table and the genome browser knows how to access each set of data per chromosome. Take a look at example table names in the hg18 database. Many tracks are loaded like this. Even if you tried to encode this file with bedToBigBed you would still need a very large machine. Probably even larger than the hgLoadBed command requires. --Hiram Jennifer Jackson wrote: > Hello, > > It seems like you are using your own mirror and trying to upload there as a > native track? From a first look, you are out of disc space. > > Double check that you have space, then perhaps try an alternate format > designed for large datasets, such as bigBed or bigWig. > > And finally, breaking the data up per-chromosome is always an option. We do > this for many tracks internally. > > Hiram may have more to add later on, > > Thanks, > Jennifer > > ps. Just to double check, your mySQL version is still 4, but we are on 5. Be > certain that you are using a consistent source code tree for all (data and > software). Some changes were made to several utilities when the switch to v5 > was made. See the genome-mirror mailing list archives for exact changes. > http://genome.ucsc.edu/contacts.html#mirror > > > ------------------------------------------------ > Jennifer Jackson > UCSC Genome Bioinformatics Group > > ----- "WANG Wuqi" <[email protected]> wrote: > >> From: "WANG Wuqi" <[email protected]> >> To: "[email protected]" <[email protected]> >> Cc: "Jennifer Jackson" <[email protected]>, [email protected], "WANG Wuqi" >> <[email protected]> >> Sent: Tuesday, November 10, 2009 7:39:38 PM GMT -08:00 US/Canada Pacific >> Subject: UCSC data loading error - needMem: Out of memory >> >> Hi Hiram, >> >> When we upload BED data format with 12 columns to encounter below >> error: >> ................ >> $ hgLoadBed hg18 RHM005_final_gff_noFILT RHM005_final_gff_noFILT.bed >> Reading RHM005_final_gff_noFILT.bed >> needMem: Out of memory - request size 89 bytes, errno: 12 >> ................ >> >> The size of this data file is about 3GB. When we take part of data, >> for example, 1000 rows >> to do custom track or load, both are ok. >> >> Our server: >> Linux ucsc.fugu-sg.org 2.6.9-67.0.22.ELsmp #1 SMP Wed Jul 23 17:30:51 >> EDT 2008 i686 i686 i386 GNU/Linux >> >> MySQL: Ver 14.7 Distrib 4.1.22, for pc-linux-gnu (i686) using readline >> 4.3 >> >> >> Thanks & regards >> >> Wuqi > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
