Hi Katrina, Thank you very much for your helpful response. I did set up the kent source tree and located the perl script. I also looked up the file /kent/src/hg/makeDb/doc/hg19.txt to grep the particular command you guys used for generating the rbest chain:
/cluster/bin/scripts/doRecipBest.pl -workhorse=`uname -n` -stop=recipBest -buildDir=/hive/data/genomes/hg19/bed/lastz.$db hg19 $db >& $db.recipBest.log I scanned the perl script briefly and thought that I roughly understood the basic usage of the tool. Then I tried to do the following to construct a rbest chain of panTro3 vs rheMac2: ./doRecipBest.pl -stop=recipBest -buildDir=~/data/panTro3 ~/data/panTro3 ~/data/rheMac2 where the directory ~/data/panTro3 contains the file: panTro3.rheMac2.all.chain and the directory ~/data/rheMac2 contains the file: rheMac2.panTro3.all.chain Both are downloaded from the UCSC golden path repository. But somehow I think providing the chain files only is not sufficient for the doRecipBest script to work out. Could you please let me know what files are assumed to exist in these directories? And what is the buildDir in relation to the tDb and gDb? Thank you. Best, - Jia ________________________________________ From: Katrina Learned [[email protected]] Sent: Tuesday, August 30, 2011 3:36 PM To: Zeng, Jia Cc: Luvina Guruvadoo; [email protected] Subject: Re: [Genome] Locating orthologous regions (not just genes) between human and chimp Hi Jia, You can build the rbest.chain files yourself. You'll need to build the kent source tree and then use the doRecipBest.pl program to create them. Here is some information about where to download the kent source tree: http://genome.ucsc.edu/FAQ/FAQlicense.html#license3 Once you build the kent source tree, type doRecipBest.pl without any arguments into the prompt to see the usage, options, and dependencies: usage: doRecipBest.pl tDb qDb options: -continue step Pick up at the step where a previous run left off (some debugging and cleanup may be necessary first). step must be one of the following: recipBest, download -stop step Stop after completing the specified step. (same possible values as for -continue above) -buildDir dir Use dir instead of default /hive/data/genomes/$tDb/bed/blastz.$qDb -dbHost mach Use mach (default: hgwdev) as database server. -workhorse machine Use machine (default: least loaded) for compute or memory-intensive steps. -debug Don't actually run commands, just display them. -verbose num Set verbose level to num (default 1). -help Show detailed help and exit. Automates addition of reciprocal best chains/nets to regular chains/nets which have already been created using doBlastzChainNet.pl. Steps: recipBest: Net in both directions to get reciprocal best. download: Make a reciprocalBest subdir of the existing download dir. All work is done in the axtChain subdir of the build directory: /hive/data/genomes/$tDb/bed/blastz.$qDb unless -buildDir is given. Reading the program is the best way to understand the processing, it is not documented anywhere else. Here are some previously answered mailing list questions that may be helpful to you: https://lists.soe.ucsc.edu/pipermail/genome/2009-April/018818.html https://lists.soe.ucsc.edu/pipermail/genome/2009-April/018821.html Regarding your concern about the nets having some human chr1 <--> chimp chr15 alignments, even within humans there can be small rearrangements so it's reasonable for there to be some rearrangements between human and chimp even though at a broad scale the chromosomes are very similar. Please contact the mail list ([email protected]) again if you have any further questions. Katrina Learned UCSC Genome Bioinformatics Group On 8/26/11 11:45 AM, Zeng, Jia wrote: > Hi Luvina, > > Thank you very much for the help. The chain file is very useful to my > research. Inspired by it, I'm also interested in the rbest.chains of hg19 vs > rheMac2 and panTro3 vs rheMac2. > > I understand that you guys have a very busy schedule and I do not want to > bother too much. Is it possible for you to outline the procedure to generate > these files so I can do it by myself? > > Thank you once again. > > Best, > - Jia > ________________________________________ > From: Luvina Guruvadoo [[email protected]] > Sent: Tuesday, August 02, 2011 6:14 PM > To: Zeng, Jia > Cc: [email protected] > Subject: Re: [Genome] Locating orthologous regions (not just genes) between > human and chimp > > Hi Jia, > > The rbest.chain file for hg19 vs panTro3 is now available here: > http://hgdownload.cse.ucsc.edu/goldenPath/hg19/vsPanTro3/reciprocalBest/ > > Regards, > --- > Luvina Guruvadoo > UCSC Genome Bioinformatics Group > > > On 6/28/2011 10:24 AM, Luvina Guruvadoo wrote: > > Hi Jia, > > We have noted your request and will notify you once the rbest.chain file > for hg19 vs panTro3 has been released. > > Regards, > --- > Luvina Guruvadoo > UCSC Genome Bioinformatics Group > > > Zeng, Jia wrote: > > > Hi, > > I have this question and after searching the mailing-list archive, I found a > similar thread for which a UCSC staff member Jennifer gave a very good answer > to: (https://lists.soe.ucsc.edu/pipermail/genome/2010-January/021123.html) > > But my question is a bit more general than the original post so I would > appreciate a bit more information. > > I need to perform a genome-wide search to locate all the orthologous regions > (including both non-coding and coding regions) in the human genome that are > orthologous to the chimp genome. From the original answer in the related > post, I have an impression that to do so, the best way is to get a hold of > the "best reciprocal alignment" between the human and chimp genome instead of > just the single directional alignment such as the files shown in > http://hgdownload.cse.ucsc.edu/goldenPath/hg19/vsPanTro3/axtNet/ (I kind > of presume the single directional alignment does not necessarily imply > orthology since some regions in the human chromosome 1 correspond to the > chimp chromosome 15, which doesn't make sense). > > However I failed to locate the rbest.chain file in the newest human/chimp > assemblies, which used to exist in earlier assembly (eg. hg17 vs panTro1). > Could you please let me know how I can construct the best reciprocal > alignment between human and chimp using two single directional alignments > (human vs chimp and chimp vs human)? Can I just naively intersect the two > single directional alignments and retain the entries that exist in both > files? Please kindly advise. Thank you very much for your time in advance. > > Best, > - Jia > _______________________________________________ > Genome maillist - [email protected]<mailto:[email protected]> > https://lists.soe.ucsc.edu/mailman/listinfo/genome > > > > > _______________________________________________ > Genome maillist - [email protected]<mailto:[email protected]> > https://lists.soe.ucsc.edu/mailman/listinfo/genome > > > > _______________________________________________ > Genome maillist - [email protected] > https://lists.soe.ucsc.edu/mailman/listinfo/genome _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
