Help Downloading Mouse Mm9 Refseq Gene Coordinates

Question

[komal.rathi]

Hello,

I am making a bed file of mouse mm9 coordinates for all the genes expressed in the organism. For this purpose, I want to download all the RefSeq genes' coordinates for mouse mm9 assembly. Any suggestions?

[Ashutosh Pandey]

Did you try searching for similar questions? I assure you that it has been answered several times.

[komal.rathi]

I tried downloading RefSeq genes from the UCSC table browser but I have a list of test genes and they are not present in the data that I downloaded. I am looking for something more comprehensive.

[Jelena Aleksic]

Where's your list of test genes from? Is it from RNAseq or something? Which genes are breaking?

[komal.rathi]

I have a list of genes with coordinates from Affy mm10 2.0 annotation. My initial goal was to convert the coordinates to mm9 using LiftOver. But after using LiftOver, when I checked a couple of genes in the UCSC genome browser, the coordinates were not correct for the genes. So I am trying to download RefSeq genes for mm9 assembly and get the coordinates by matching the RefSeq IDs.

[Devon Ryan]

What are a few of the names that were missing?

[komal.rathi]

I couldn't find NR_051982 which is for Dennd2d gene. I downloaded the data from UCSC table browser, track: RefSeq Genes, table:refGene.

[Devon Ryan]

It probably doesn't help that NR_051982 was replaced with NM_001278941.

Answer 1

[Jelena Aleksic]

Based on your comments: I think you might be confused about genome releases and annotation releases. Genome releases are relatively infrequent - mm9 was in July 2007, while mm10 happened in 2011. On the other hand, annotation releases are quite frequent - every few months or so. The difference is that the genome releases change the underlying coordinate system, while the annotation releases update the gene models. So even when you're working in the same genome release (e.g. mm10), the location / existence / classification of some genes can still change every few months or so. Taking a current annotation release and mapping it back might give you accurate relative coordinates, but might not match the correct genes, because the gene models used might have been different at the time. You're likely to run into some problems with RefSeqIDs too, because both gene models and IDs change over time. However, I'm not sure why you're translating your coordinates back to mm9, rather than updating everything to mm10.

[komal.rathi]

We have a lot of data from the Encode Mouse Experiment matrix which is only available in mm9. So instead of converting everything to mm10 we thought we'd convert just one dataset to mm9.