Protein Alignment Tool

I found a bug in my build.xml script. I let Eclipse compile all my source with the exception of Version.java which is dynamically built by Ant and then compiled before I generate the project JAR file. The problem is that even though my Eclipse project settings were set to 5.0, I didn’t set the target in the javac task in the build file. This resulted in a Bad Version Number error if I built the project on a machine where Java 6 was the default.

This is a problem I ran into when adding activation.jar and mail.jar (both from Sun) to my .jnlp file. It turns out that these JAR files are already signed by Sun. All JAR files within a JNLP file must be signed by the same certificate - the way around it is to have multiple JNLP files and use the extension keyword. For details see this post from Kirill Grouchnikov’s Blog.

Protein sequences stored in UniProt have a CRC64 checksum.  This means that it is possible to take an arbitrary sequence (with or without a known identifier), calculate the CRC64 checksum for it, and then search UniProt for that specific sequence.  There are two problems with this:

1. UniProt does not provide any web services for querying their database.  I’ll check again, but, I don’t see it
2. There is a low possibility of collision; that is, two sequences may produce the same CRC64 checksum.

On Nature Precedings I came across A SEquence Globally Unique IDentifier (SEGUID) Proteome Database that solves the two problems above.

Powered by ScribeFire.