Hello, The liftOver version in the Downloads area is slightly different than the one used in hgConvert (the Browser based liftOver tool). But not this different.
But my guess about the source of the discrepancy was the variable -minMatch. The default in the line command tool is 0.95 (strict, intended for lifts from one version of the same genome to another version of the same genome). The default in the Browser tool is 0.1 to allow for cross-species lifts to have a better chance of success. Testing proves this is the problem. I would change the Galaxy default for liftOver to use -minMatch 0.1. It really cannot hurt if you do not use -multiple. With the example data, I got the exact same results using the Browser tool as the user: 2496 successful conversions and 1265 failures. When I run the line command of liftOver and specify -minMatch 0.1, I get exact results: 2496 successful conversions and 1265 failures (note that with the "unmapped" file there are two output lines per line of input that fails, one line is the description of why the failure occurred and the other is the data point that failed). When I run the line command liftOver and and do not specify -minMatch (therefore using the default of 0.95), I get: 1651 successful conversions and 2110 failures. Glad I could help, please let us know if you need more info, Jennifer Jackson Guruprasad Ananda wrote: > Hi, > > I'm forwarding an email from a Galaxy user, which points out a > possible discrepancy in the outputs obtained by using liftOver binary > downloaded from UCSC versus liftOver utility available on UCSC > website. There is also a sample input file attached, which produces > different outputs in the 2 cases. Can you please advise us on why > this could be happening? > > Thanks, > Guru. > Galaxy Team. > > > Begin forwarded message: > > >> From: "GALIH Kunarso" <[email protected]> >> Date: March 25, 2009 5:11:12 AM EDT >> To: <[email protected]> >> Subject: [galaxy-bugs] liftOver discrepancy >> >> Hi Galaxy, >> >> Firstly, thanks for your nice website, it’s very useful and I use >> it a lot! >> >> I would like to highlight to you here a difference between the >> liftOver results from your main site to the one I got from the UCSC >> site. >> >> For example, I attach here 100bp-long regions surrounding the Oct4 >> binding site identified by ChIP-Seq in mouse Embryonic Stem cell >> (Chen X, Xu H, et al. Integration of external signaling pathways >> with the core transcriptional network in embryonic stem cells. >> Cell. 2008 Jun 13;133(6):1106-17). >> >> When I liftOver-ed this using the main Galaxy site from mm8 to hg18 >> I got 1651 successful conversions and 2110 Unmapped coordinates. >> When I tried this on UCSC site I got back 2496 successful >> conversions and 1265 failures. I just used the default parameter of >> the UCSC site for this (Minimum ratio of bases that must remap: >> 0.1, Minimum chain size in target: 0, Minimum hit size in query: 0, >> Allow multiple output regions: Unticked, Min ratio of alignment >> blocks/exons that must map: 1, If thickStart/thickEnd is not >> mapped, use the closest mapped base: Unticked). >> >> Based on my simple checking, some of Galaxy’s “#Partially >> deleted in new” are considered as success by UCSC. >> >> Can you help me regarding this? Is your liftOver parameter >> different from the default UCSC one? >> >> Thank you for your attention. >> >> >> Regards, >> >> Galih Kunarso >> Computational and Mathematical Biology >> Genome Institute of Singapore >> 60 Biopolis Street, Genome #02-01, Singapore 138672 >> DID: 6478-8661 Fax: 6478-9058 >> >> -------------------------------------------------------- >> This e-mail is confidential and may be privileged. If you >> are not the intended recipient, please delete it and notify >> us immediately. Please do not copy or use it for any purpose, >> or disclose its contents to any other person. Thank you. >> -------------------------------------------------------- >> >> >  > >> _______________________________________________ >> galaxy-bugs mailing list >> [email protected] >> http://mail.bx.psu.edu/cgi-bin/mailman/listinfo/galaxy-bugs >> > > Regards, > > Guruprasad Ananda > Graduate Student > Bioinformatics and Genomics > The Pennsylvania State University > > > > _______________________________________________ > Genome maillist - [email protected] > http://www.soe.ucsc.edu/mailman/listinfo/genome > _______________________________________________ Genome maillist - [email protected] http://www.soe.ucsc.edu/mailman/listinfo/genome
