Hello Dan, The phastCons data was generated with an older version of the phastCons software (probably version v0.9.9.6b from late 2006 or early 2007) and may be the cause of the differences you are seeing. You might want to consider using the 46-way data instead from hg19 available here:
http://hgdownload.cse.ucsc.edu/goldenPath/hg19/phastCons46way/ If you have further questions or require clarification feel free to contact the mailing list at [email protected]. Regards, Pauline Fujita, Ph.D. UCSC Genome Bioinformatics Group http://genome.ucsc.edu On 09/28/11 07:32, Daniel Halligan wrote: > Hi, > > I'm attempting to reproduce the phastCons most conserved elements tract for > placental mammals from the multiz-28way alignments. I've initially attempted > to do this just for chromosome 1, and whilst there is fairly good agreement, > I'm getting some differences. > > The phastCons command I've used is: > phastCons -i MAF -n -s -E 45 -C 0.3 -R 0.31 --not-informative panTro2,rheMac2 > -V chr1.bed multiz28way/maf/chr1.maf placental.mod > > where placental.mod is taken from here: > http://hgdownload.cse.ucsc.edu/goldenPath/hg18/phastCons28way/placental.mod > > and is as follows > ALPHABET: A C G T > ORDER: 0 > SUBST_MOD: REV > BACKGROUND: 0.295000 0.205000 0.205000 0.295000 > RATE_MAT: > -0.990634 0.178819 0.490738 0.321078 > 0.257325 -1.001677 0.186563 0.557790 > 0.706184 0.186563 -1.161873 0.269126 > 0.321078 0.387615 0.187019 -0.895713 > TREE: > (((((((hg18:0.005873,panTro2:0.007668):0.026074,rheMac2:0.031973):0.073300,otoGar1:0.151185):0.015682,tupBel1:0.162844):0.006272,(((rn4:0.084383,mm8:0.076274):0.200607,cavPor2:0.202990):0.034350,oryCun1:0.208548):0.014587):0.019763,((sorAra1:0.248532,eriEur1:0.222255):0.045693,(((canFam2:0.101137,felCat3:0.098203):0.048213,equCab1:0.099323):0.007287,bosTau3:0.163945):0.012398):0.018928):0.030081,(dasNov1:0.133274,(loxAfr1:0.103030,echTel1:0.232706):0.049511):0.008424); > > > The majority of elements have identical coordinates, although a large number > differ by a single base (and in these cases the predicted conserved elements > that I download from the UCSC genome browser start 1bp before the conserved > elements I find using the above command). > > Furthermore, the predicted conserved elements I obtain contain a single very > large element amongst the other predictions which has a very low (and > negative) score. In BED format, this element is: > chr1 120376987 247198473 chr1.105012 -999999999 + > > I'm not sure why I am observing these differences, but I would be very > grateful if you could help explain what might be going on. > > Many thanks, > Dan Halligan > > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
