Why is the start position always shifted one base pair?

43 views
Skip to first unread message

Emma

unread,
Apr 9, 2026, 9:30:29 AMApr 9
to UCSC Genome Browser Public Support
When I try to lift from hg38 to hg19 I get that the reference coordinates in hg38 are not exactly the same as my input, why is this?

So for example if I lift: 
chr1 898703 6229913

the result is:
chr1 834083 6289973 chr1:898704-6229913 1

Why is 898704 displayed and not 898703?
The end position seems to be the same though. 

Thanks for your help!


Emma

unread,
Apr 9, 2026, 10:30:32 AMApr 9
to UCSC Genome Browser Public Support, Emma
Never mind, I think it's giving the coordinates in "position box" format instead of BED format. 

Emma

unread,
Apr 9, 2026, 11:30:38 AMApr 9
to UCSC Genome Browser Public Support, Emma
Is there a reason why -preserveInput doesn't work with position box format? Is there a way to force it?

Gerardo Perez

unread,
Apr 15, 2026, 10:31:19 PM (14 days ago) Apr 15
to Emma, UCSC Genome Browser Public Support

Hello, Emma.

Thank you for your interest in the UCSC Genome Browser and for sending your inquiry.

The difference you noticed between 898703 and 898704 is due to coordinate systems. BED format uses 0-based starts, while position format uses 1-based starts. This blog post has a great explanation of how coordinates work in the two systems: https://genome-blog.gi.ucsc.edu/blog/2016/12/12/the-ucsc-genome-browser-coordinate-counting-systems/

Are you using the command-line liftOver utility or the web LiftOver? In your example, it looks like you ran liftOver for a BED3 input:
chr1 898703 6229913. The -preserveInput setting adds the original input coordinates to the output name column (column 4) in position format (1-based). There is no current setting for it to display in BED format; it was designed to display in position format.

I hope this is helpful. Please include gen...@soe.ucsc.edu in any replies to ensure visibility by the team. 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.

Gerardo Perez
UCSC Genomics Institute


--

---
You received this message because you are subscribed to the Google Groups "UCSC Genome Browser Public Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genome+un...@soe.ucsc.edu.
To view this discussion visit https://groups.google.com/a/soe.ucsc.edu/d/msgid/genome/7d379775-6b0a-431a-a532-9c442a915b0fn%40soe.ucsc.edu.
Reply all
Reply to author
Forward
0 new messages