HG38 to HG19 liftover question

[GLEB KICHAEV]

Hello,

I am attempting to map HG38 coordinates to HG19 coordinates using the liftOver tool. It seems to work correctly apart from the fact that on occasion, its map regions to completely different chromosomes. For example,

HG38:

chr1 450740 450742

maps to

HG19

chr5 180795223 180795225

I just wanted to make sure that this was correct and that the utility is working as intended. I believe I am running it correctly. I use the following command to execute the program:

liftOver GenCode.UTR hg38ToHg19.over.chain GenCode.UTR.hg19 GenCode.UTR.unmapped

where "GenCode.UTR" is just a standard bed file.

Many thanks for your help,

Gleb Kichaev 

[Matthew Speir]

Hi Gleb,

Thank you for your question about lifting coordinates from hg38 to hg19. If you are ever unsure about liftOver results, I recommend using the Genome Browser to look at the region and the sequence in detail. You can use the Genome Browser to grab a 20bp region that includes your 2 bases of interest from hg38. Here's the sequence I used:

    >hg38_dna range=chr1:450732-450751
    TATATCATTTATGAGATCCT

If you use BLAT, http://genome.ucsc.edu/cgi-bin/hgBlat?db=hg38, to map that sequence to hg19 and hg38, you'll notice that there are hits to a few different positions in the Genome, all with 100% identity. It appears that the sequence you're interested in occurs in a few different positions in the genome, and when liftOver outputs these positions, the chr5 position is output first. You can use the "-multiple" option for liftOver to output the other two positions.

I hope this is helpful. If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible Google Groups forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.

Matthew Speir
UCSC Genome Bioinformatics Group

--