====== ENCprime =======

A program that calculates a codon usage bias summary statistic, Nc'. It is based
on the effective number of codons statistic Nc (or ENC) developed by Frank
Wright, but improves upon it by accounting for background nucleotide
composition. The package includes SeqCount, a companion program that prepares
FASTA format files for use by ENCprime.

These programs are all distributed free of charge. Please contact me with any
problems you have with these programs whether it be bugs or built-in
limitations.

For full user documentation, consult ENCprimedoc.pdf

===Note to users===

The team at bioinfo.hr has made an R package that computes Nc' that 
may be more accessible/useful to modern users: https://github.com/BioinfoHR/coRdon

I say this as the C code here written in the summer of 2001 is showing its age and 
portability issues seem to have arisen.  In particular, there is an unresolved bug 
in SeqCount where it is miscounting the sequence length by 2 on at least some 
systems, so I advise testing with a small test sample to make sure it is working well
on your system. 

=== UNIX/Linux distribution ===

For installation, type:

make all

If you have problems, it may be because 'gcc' is not a compiler that is
available on your system.  If so, find out what is the standard
compiler, and adjust the makefile accordingly.

The binary executable is configured to be output in the 'bin' subdirectory.

Note: The Unix version runs under Mac OSX.

=== Win32 version ===

For Windows users, Anders Fuglsang
has kindly provided a zip archive of ENCprime compiled on a Win32 XP system.
Windows users can also try Fran Supek's Windows-based INCA package for codon
usage analysis which computes ENCprime among other statistics. As a caution,
neither program has been extensively tested by this author. Furthermore, Forrest
Zhang has reported some problems with the Win32 version of SeqCount garbling
sequence names (6/2/06).

=== Development notes ===

Development Notes:

1/12/14 Moved the source into a github repository.  On occasion users have reported issues with very large files crashing the program.  A TODO item is to recheck and revamp as necessary the memory allocation calls.

2/28/06 Anders Fuglsang found a bug in the calculation of the 3-fold average
homozygozity when no 3-fold redundant codons are observed. In this case, the
3-fold average homozygosity is supposed to be estimated by taking the average of
the 2-fold and 4-fold average homozygosities. The original code took the
harmonic rather than arithmetic mean. The bug should have had only a small
effect on datasets for which no 3-fold redudant codons were observed. The
revised code contains comments showing the bug.

8/23/04 The documentation has been updated to address how ENCprime calculates Nc
and Nc' when a small number of codons of any particular amino acid has been
observed. The text helps explain why calculation of Nc by other programs may
result in different values for short sequences. The basic reason is that
different programs to compute Nc use different corrections for when there is
insufficient data.

7/21/03 Multiple users have noticed SeqCount will crash with large files. This
appears to happen during a calloc call within the program. The source of this
bug is being investigated. For now, the crashing can be avoided by dividing
large data files into smaller files and running them in batch.

1/20/03 Mike Cummings found a bug that caused SeqCount to crash when run with
only a single sequence. The error had to do with how files were being closed,
and has been fixed. A small bug was also fixed that caused ENCprime to crash
under MacOSX.

1/8/03 Mike Cummings helped find a bug in the interactive mode of ENCprime. The
genetic code setting would not change appropriately. This bug has been fixed.

9/11/02 Some small changes to the documentation were made.

9/3/02 A bug in SeqCount's calculation of the nucleotide composition was fixed.



If you have other problems, consult your local compiling guru, or
e-mail jnovembre@gmail.com.