Sunday, May 24, 2009

Promethease 0.1.66 is out

As always there are many improvements since the last blog post, but these standout.

There is a new syntax for genosets. Older clients don't know how to read this, and will not be able to see these genosets. This is a good reason for you to upgrade to the latest version.

Since 0.1.50 it has been possible to pay $2 via and reduce the runtime to about 5 minutes. In doing so you're downloading a single compressed cachefile instead of making thousands of requests for individual pages. At the moment this cache file is updated approximately weekly, however in the future we expect to have this updated in realtime as the wiki is edited. There will be a new blog post when that works. Until then the $2 version is faster, but the free version uses slightly fresher data.

The upper right corner has a new button labeled 'Wizard'. Press this to use a simpler UI which walks you through setting up your run. If you pay the $2, not only will you get the cachefile for a faster run, but you'll also have access to 2 new features with this wizard. The first is the ability to only show pages which have been modified recently. This might be useful if you ran your report 2 months ago, and want to make a new report which only shows what is new since then.

The wizard also enables a highly experimental new kind of report. This report is based on loading multiple genotype files and predicting what is possible in the offspring. You might wish to use this if you have data for you and your spouse, and you'd like to see some possible genotypes in your children. Its really not quite ready for the general public, but it never will be unless it is released for wider testing. Most folks should just click 'Next' and move on to generate a traditional report.


jf said...

Wow! This worked great - just under 1 min processing time for my 23andMe raw data with the Amazon option. Thanks for putting together such an amazing tool!

Unknown said...

glad to hear it. I'm curious to hear runtimes for other people, since that is a bit faster than I'd expected. The last line of the log should actually tell you runtime.

jf said...

I re-ran the analysis but this time did a "breeding" comparison with my wife's genome. Log says 214 seconds - amazing!

John said...

I dwlded it, but its not working.

Unknown said...

Windows? Mac? Windows7? shows nothing? Shows but won't run? Leaves a snpedia.log file? How big is your download?

John said...

I replaced my old version with this new one and when I hit it, it shows me an option to run, but after hitting run, its not doing anything.

Unknown said...

Now we're getting somewhere. Your last reply didn't indicate an answer to my Mac vs Windows question. However Affy GCOS sheds some light.

Promethease reads the formats I've seen. I have decent support for the Navigenics format. A sample one is at
If you can make your file look like that (and obey all the same orientation rules) you're set. Feel free to use a gziped version.

Promethease's support for Navigenics is decent. Its not based on as much testing as the 23andMe platform, so there may still be some flipping issues. However I've encountered several other data sets from the same Affy 6 platform, in different formats. It seems it doesn't yet understand yours. I'm afraid without a sample I won't much be able to help. If that is something you'd like to discuss, but not on a blog comment I can be reached as

John said...

Cariaso, I am using the data, but just trying to start the program. Its working in my other computer, but not the one I use in my office. I guess there is some other issue. Thanks for all the replies. I will email u if I need further clarification on Affy Data. Thanks.

MEJ said...
This comment has been removed by the author.
MEJ said...

"Promethease's support for Navigenics is decent...."

Can anyone walk me through downloading raw data from the Navigenics web site?

I can not seem to locate it on my own and I assume support is off for the long weekend!