[Biojava-dev] Fwd: [Biojava-l] Port an Algorithm to Java

[Andreas Prlic]

I think it really depends on the use case. I recently added FATCAT and
CE, two prominent protein structure alignment algorithms to the
biojava-structure module. While doing so some caching layers and
support for multi-threaded and distributed calculations were added.
Also standard Java tricks like allowing objects to be serializable,
using beans etc. were useful additions. The end product allowed me to
run 1 billion protein structure alignments on jobs spread across
multiple datacenters and heterogenous hardware architectures. This
would not have been straightforward at all using the original C
code...

Andreas



On Tue, Apr 3, 2012 at 8:01 PM, Russ Kepler <russ at kepler-eng.com> wrote:
> On Wednesday, April 04, 2012 11:52:18 AM Smithies, Russell wrote:
>> Another option
>> if you really want to integrate blast+ might be JNI?
>
> Back when I was doing a lot of that it worked out as well to integrate with
> the command line executable - the output files were generally transportable and
> often could be parsed from the standard output.  I don't think it would work
> as well with an application with fine granularity but those aren't very common
> in this line of work.
>
> _______________________________________________
> biojava-dev mailing list
> biojava-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biojava-dev