Please note, bugs in phastCons have been fixed since those data were produced.
You can find updated predictions without this off-by-one bug in:
ftp://hgdownload.cse.ucsc.edu/goldenPath/hg19/phastCons46way/
We are not going back to earlier data to rework it.
--Hiram
Daniel Gaffney wrote:
> Hi Malcolm,
> Thanks for the update - does it seem odd to you that both our examples show
> that problems occur in the lines preceding a block
> whose start coordinate is "....00001"? Could this be an 0/1 error in array
> indexing perhaps?
> Best
> Daniel
>
> On Wed, Sep 28, 2011 at 7:40 PM, Cook, Malcolm <[email protected]> wrote:
>
>> +1: http://www.mail-archive.com/[email protected]/msg03607.html
>>
>>
>> -----Original Message-----
>> From: [email protected] [mailto:[email protected]] On
>> Behalf Of daniel.gaffney
>> Sent: Wednesday, September 28, 2011 11:01 AM
>> To: [email protected]
>> Subject: [Genome] PhyloP off by one errors?
>>
>> Hi,
>> I've been trying to convert the hg18 phyloP scores for placental mammals to
>> bigWig format and I've come across what look
>> like numerous off-by-one errors (and others) in the files posted here:
>>
>> ftp://hgdownload.cse.ucsc.edu/goldenPath/hg18/phyloP44way/placentalMammals/
>>
>> I've been using wigToBigWig from the UCSC source to do the conversion. I
>> use the following code:
>>
>> zcat $FILE | wigToBigWig -clip stdin chromSizes.txt test.bw
>>
>> where FILE is a wig file from the above directory. So far I've found the
>> following errors:
>>
>> There's more than one value for chr1 base 220000001 (in coordinates that
>> start with 1).
>> There's more than one value for chr13 base 60000001 (in coordinates that
>> start with 1).
>> There's more than one value for chr16 base 20000001 (in coordinates that
>> start with 1).
>> There's more than one value for chr2 base 60000001 (in coordinates that
>> start with 1).
>> There's more than one value for chr3 base 180000001 (in coordinates that
>> start with 1).
>> There's more than one value for chr4 base 110000001 (in coordinates that
>> start with 1).
>> There's more than one value for chrY base 20000001 (in coordinates that
>> start with 1).
>>
>> Removing the value line before the wig line e.g.:
>>
>> 0.561 <==== delete this line
>> fixedStep chrom=chr1 start=220000001 step=1
>>
>> seems to fix the problem, but its pretty time-consuming. I hope this is of
>> use to somebody - but any ideas why its happening?
>> Best
>> Daniel
_______________________________________________
Genome maillist - [email protected]
https://lists.soe.ucsc.edu/mailman/listinfo/genome
