Skip to main content

Frequently Asked Questions

FAQ. 2010-06-29: Why do so few methods have a help() page?

A: The main source of documentation for the aroma framework, is the http://www.aroma-project.org/ website. What is documented there, is well maintained, tested and supported. The help() pages available that comes with the different aroma.* packages are mainly for developers and much less to for end users. Several methods and classes do not even have such help pages. The main reason is lack of resources to write but also maintain such documentation. Please consider contributing to the online as well as help-page documentation.

FAQ. 2008-03-26: Why do you not want to answer questions sent to your private email address?

A: Please post all questions related to aroma.affymetrix to its dedicated mailing list. This way there is a chance that others also can help out with your problem/question, which means that you hopefully get a reply quicker. All questions and answers are archived and searchable online. Because of this, we can then refer to already answered questions by replying with a link. Moreover, this way the chances for finding answers online increase. So, please understand this when I ask you to repost your question to the mailing list. It will not only save me but everyone a lot of time.

FAQ. 2007-07-13: Hmm... why does R crash (core dump)?

A: This is most likely due to a corrupt CDF or CEL file that the affxparser package is trying to read, e.g. zero-size or incomplete files. The solution is to turn on as much verbose output as possible so that the corrupt file can be located and removed. Unfortunately we cannot control for this at the moment, because it is due to the underlying Fusion SDK native library that affxparser is using, not R or aroma.affymetrix. Hopefully it will be more robust in the future. However, we have added some sanity checks to avoid the most obvious corrupt files.

FAQ. 2007-07-09: How do I convert an ASCII CDF into a binary CDF?

A: Working with binary CDFs is much faster and much more memory efficient. You can convert an ASCII CDF into a binary CDF using the convertCdf() function in affxparser.

FAQ. 2007-05-24: Where can I download CDF files

A: The CDF file for a given chip type is available in so called "Library Files" at the corresponding "Support Materials" page on the Affymetrix website. You may find links to those Affymetrix pages via the Chip types pages.

FAQ. 2007-03-07: Where is the file cache located?

A: The call getCacheRootPath() will tell you where the root directory of the file cache is located on your system. All of the cache files created by the package is under subdirectory aroma.affymetrix/. You can use setCacheRootPath() to change the root directory.

FAQ. 2007-02-14: The NnnExplorer does not seem to work?

A: The NnnExplorer:s, such as ChromosomeExplorer and ArrayExplorer, are developed for modern web browser. If it does not work on your browser, make sure to update and if that does not help, please let us know.

FAQ. 2007-02-13: Why is aroma.affymetrix slow?

A: The package is not slow ;) Seriously, it is designed to deal with any-sized data sets by keeping data on file rather than in memory. In other words, we do pay a price from I/O of data. For most algorithms this means that data is read twice instead of once, but that extra overhead is not large. However, you will experience that it takes a long time when you process a new chip type for the first time. This is because the package needs to create a customized CDF file ("monocell CDF"), which takes time. As soon as this is created, you should not expect to see any longer delays. Moreover, to improve on speed, the package caches (to file or in memory) results from calls that are computational expensive and are likely to be done again, either in the same session or in a future session. You will see that this caching further speeds things up. Also, make sure to use the verbose arguments to see what is going on when the package appears to "get stuck". Update 2008-11-04: See also the 'Improving processing time' how to page.

FAQ. 2007-02-09: What is a monocell CDF?

A: A monocell CDF is a CDF that has been extracted automatically from the "main" CDF and it can be identified by its name *,monocell.cdf. It only contains one cell per unit group. Monocell CDFs are used to store chip-effect estimates of two-way PLMs. A monocell CDF is created automatically if missing. This may take everything from a few minutes to an hour. However, is only done once per chip type.