UCSC genome browser negative strand positions

Question

[ijarne]

I have been working on some data recently and we have encountered a very curious behaiviour of UCSC genome browser. We have been working. We have been checking deletions and duplications that happen in some of our samples. When make a transformation where we pick the mutation and then we transform their coordinates to the coordinates of the exons that is occupies for the sake of better annotation in the future.

The thing is that when we do this on the positive strand of the genome we can check in the browser and our coordinates match perfectly. But when we do the same on the negative strand genes, we find our coordinates to be between the genes that we are wantig to check. Any suggestions? Is this a common thing ?

So for example when we have a mutation between exons 5 and 12 of gene AMN, which is in the + strand. I then take the start and end of those exons I can can see them perfectly.

But if I have a mutation between exons 1 and 2 of gene MECP2, in the - strand. When I enter that data in the UCSC, I am getting this.

As you can see my positions map perfectly to the intron between those exons. Could anybody help me understand how this is happening ? Thanks!

[Pierre Lindenbaum]

hen make a transformation where we pick the mutation and then we transform their coordinates to the coordinates of the exons that is occupies for the sake of better annotation in the future.

isn't it a problem with the tool use to convert the coordinates ?

[rfran010]

Would this be related to some function that takes the "end" as the start for negative strand features?

E.g.

Exon1: chr1 100 500
Exon2: chr1 700 800

It seems you want the coordinates: chr1 100 800 which would display the leftmost of Exon1 and rightmost of Exon2. On plus strand features, there is no confusion as the start and ends are always consistent with starts on the left and ends on the right. However, on negative strand features, we often take the rightmost column as the start (for example, the TSS of a negative strand gene would be the rightmost value)

So applied to the example, this would provide the "start" of Exon1 as 500 and "end" of Exon2 as 700. This would effectively be supplied as chr1 500 700, and only display the intron. Also note, the Exon1 and 2 designation would be switched from the example.

Maybe another possibility, if Exon1 and Exon2 are not handled appropriately on negative strand.

+ strand
Exon1: chr1 100 500
Exon2: chr1 700 800

- strand
Exon2: chr1 100 500
Exon1: chr1 700 800

Here, if you took the leftmost start of Exon1 and rightmost end of Exon2, it would display as expected if gene is on the + strand (chr1 100 800), but not if gene is on the - strand (chr1 500 700). Although I feel this is a harder mistake to make...

Answer 1

[Maximilian Haeussler]

I don't understand the question: this has nothing to do with the browser, right? You use a tool to transform coordinates and that tool gives you strange coordinates, but the browser is only your tool to display the data.