Navigation Banner
 

     Summary: Given a nucleotide, aminoacid or codon data file without deletions, UpperBound calculates the upper bound on the likelihood value achievable by most ML methods, namely the methods which assume the distribution of column types to be a particular case of the multinomial distribution. The value reported is the ML for the multinomial distribution:

     where N is the number of distinct column types, f is the number of columns of a given type, and p is the relative frequency of columns of a given type.

     Input: A nucleotide, aminoacid or codon data file in any recognizable format. HYPHY uses the following table to translate nucleotide ambiguities (or aminoacid characters). For codon files, any of the predefined genetic code translation tables can be used to interpret the data.
    The upper bound is valid for data without any ambiguities.
    The user will be given two swapping strategies to choose from:

    Models: N/A

    Output: The value of the upper bound. For example:

Upper log-likelihood bound on the data set:
HAL 9000:Programming:DNAProject:data:hiv.seq
Sequences : 13
Sites     : 273
Site types: 85
Upper bound on likelihood = -783.455  

     Result Processing Tools: None available.

 
Sergei L. Kosakovsky Pond and Spencer V. Muse, 1997-2002