acd files

[Peter Rice]

calvin wang wrote:
> IHow to include new methods in emboss? just by
> writting the corresponding acd file? one the acd file
> be written are there modifications to be done over the
> source code of the implemented method? is there a good
> example about how to writte acd files so new methods
> can be part of emboss? 

You need to do the following:

1. Write the ACD file to read the input you need, open the output file, 
and provide any other options

2. Test it using acdc

3. Either:

3a. Write a program in C to use the ACD file (with an embInit call) and 
to rnu the algorithm or launch the application with a system call (that 
is how emma runs clustalw for example)

or

3b. modify the original program (if it is in C or C++) to read from the 
ACD file and put the parameters where the original program stores them 
(that is how phylip was done, and how I am currently converting phylip 3.6)

> The method I would like to include is TCoffee, so I
> may have to writte the acd file and modify the sorce
> code of TCoffee, right? 

That has a bioperl interface too, I believe. You could launch the 
original or the bioperl version (see emma, and remember that has a very 
complicated ACD file - life is normally much easier), or as tCoffee is 
in C you could modify the get_cl_param function in tCoffee's util.c 
file, or (maybe better) replace the calls to cl_get_param by ajAcdGet calls.

Plenty of choices. If you have problems, you can mail 
emboss-bug at embnet.org and we wil. be very happy to help.

And yes, we do need a good example of how to do this. I will try to put 
something together (perhaps a little perl script and a little C program 
as example applications - we did something similar on an ACD course here 
at EBI earlier this year)

regards,

Peter Rice