[BiO BB] multinomial distribution and Ig light chains

A.J. Rossini rossini at blindglobe.net
Wed Apr 9 12:25:20 EDT 2003

Maurya Kaut <mkaut at bu.edu> writes:

> I'm attempting to replicate the analysis of immunoglobulin genes as in
> this article:
> Lossos, et al.  The Inference of Antigen Selection on Ig Genes.  The
> Journal of Immunology 165(9): 5122-5126 (2000)
> The article is available here:
> http://www.jimmunol.org/cgi/content/full/165/9/5122
> The Java applet mentioned in the paper is here:
> http://www-stat.stanford.edu/immunoglobin/
> The contact information in the paper no longer appears to be
> valid. Basically, I'd like to understand the multinomial tail

Both Rob and Naras are still at Stanford Stat.

> having trouble putting it all together.  I've written a Perl script
> that calculates expected replacement frequency for Ig light chain
> germline genes with some success, but the numbers I get for P values
> are two or three orders of magnitude off.  

Sounds like an implementation error -- I doubt if the distribution is
pathological enough to admit to round-off problems on that order of

> Firstly, I would just like to know if there is anyone who is
> familiar with this type of statistical work.  Also, I've heard of
> the "S" engine, and its cousin "R", but I'm not quite sure if they
> are applicable here.  Has anyone used them in conjuction with
> Perl/CGI?  Any advice is appreciated.

The S statistical programming language, as implemented by S (not
generally available), S-PLUS (commercially available) and R
(open-source), is a full featured language for programming, not unlike
a functional, white-space agnostic version of Python (in a sense). 

I would (and generally only) use R for data analysis, and it'll make
programming this problem up much simpler (assuming that you know both
how to program as well as have intuition for statistical data


A.J. Rossini rossini at u.washington.edu http://software.biostat.washington.edu/ 
Biostatistics, U Washington and Fred Hutchinson Cancer Research Center 
FHCRC:Tu: 206-667-7025 (fax=4812)|Voicemail is pretty sketchy/use Email 
UW  : Th: 206-543-1044 (fax=3286)|Change last 4 digits of phone to FAX 
CONFIDENTIALITY NOTICE: This e-mail message and any attachments may be
confidential and privileged. If you received this message in error,
please destroy it and notify the sender. Thank you.

More information about the BBB mailing list