Last Updated: 11/2008 

Estimating selection intensities in overlapping genes
by Niv Sabath email:nsabath@uh.edu

The program implements a non-stationary model for nucleotide substitutions in 
overlapping genes. The program receives two aligned DNA sequences, overlaps 
orientation, overlap phase and, potentially, sequences of non-overlapping regions 
for each gene. The program fits the model to the data to infer the selection 
intensities in the two overlapping genes  

input:
seq1, seq2: two gap-less aligned homologous DNA sequences of overlapping genes.
the sequences must be in the reading-frame of the reference gene (gene A)
orientation: 1 - same-strand; 2 - opposite-strand
phase: 0 (only in opposite-strand overlaps), 1, or 2
non_ov_a1, non_ov_a2 (optional): two aligned homologous DNA sequences of the non-overlapping
regions of the reference gene (gene A). [] if not exist.  
non_b1, non_ov_b2 (optional): two aligned homologous DNA sequences of the non-overlapping
regions of gene B. [] if not exist.  
ALL sequences must be of complete codons, i.e., mod(length,3) == 0 

output:
omega_A: estimated selection intensity on gene A 
omega_B: estimated selection intensity on gene B
t: estimated time substitutions/codon 
k: estimated transition/transversion rate ratio 
like: the likelihood 
div: % divergence of seq1 and seq2 

Requires Matlab Optimization Toolbox

example:
load 'example.mat'
[omega_A omega_B t k like div] = 
   dnds_ov(seq1,seq2,orientation,phase,non_ov_a1,non_ov_a2,non_ov_b1,non_ov_b2);