codonw-slim

CodonW is a package for codon usage analysis written by John Peden in Paul Sharp's group (University of Nottingham). It was originally designed to simplify multivariate analysis of codon usage with other metrics related to codon usage calculated.

codonw-slim refactors the original codebase to add Python bindings to the underlying methods written in C and to focus on the other codon usage metrics. See below for a list of metrics and their usage.

A detailed description of each metric, with references, can be found in in the docstrings of Python methods (codonw/codonwlib/codonw.pyx). All of the multivariate analysis code has been removed since this sort of analysis is more easily done in a higher level language (e.g. FactoMineR). The interative interface has also been removed.

No error checking of nucleotide sequences is done, e.g. for start, stop codons, internal stops codons, non-translatable, and partial codons. Users should do this themself to inputs provided. For more information about how amino acids and codons have are represented internally (Recoding.md).

The source code and releases for codonw-slim can be obtained from https://www.github.com/smsaladi/codonw-slim. Please report bugs and improvements via pull requests at this repository. All modifications must pass regression testing.

Build and Installation

pip install codonw-slim

Usage

The following metrics are available:

codon adaptation index (CAI)
- CodonSeq.cai
frequency of optimal codons (Fop)
- CodonSeq.fop
codon bias index (CBI)
- CodonSeq.cbi
the effective number of codons (Nc)
- CodonSeq.enc
hydropathicity of protein
- CodonSeq.hydropathy
aromaticity score
- CodonSeq.aromaticity
Silent base composition (including GC3s)
- CodonSeq.silent_base_usage
Codon & Amino acid usage (count and relative)
- CodonSeq.codon_usage
- CodonSeq.aa_usage
- CodonSeq.rscu
- CodonSeq.raau
Base composition by codon position
- CodonSeq.bases
Base composition in all frames, Length of gene, Number of synonymous codons, G+C content (overall and by codon position), G+C content of synonymous codons at the 3rd position, G+C content of non-synonymous codons at the 3rd position, Number of synonymous codons, Number of amino acids
- CodonSeq.bases2
Dinucleotide count by frame
- CodonSeq.dinuc

As written above, each is a method of the codonw.CononSeq object, e.g.

import codonw
cseq = codonw.CodonSeq("ATGAATATGCTCATTGTCGGTAGAGTTGTTGCTAGTGTTGGGGGAAGCGGACTTCAAACG")
cseq.cai()

The return type can be a simple value, pd.Series, or pd.DataFrame.

The genetic codes can be specified by setting the CodonSeq.genetic_code property with a pd.Series whose index is a codon and value is the single letter amino acid. Instantiate an object and see CodonSeq.genetic_code for more details.

Some indicies have an option of reference values to choose from (e.g. CodonSeq.fop). Several references values can be chosen by specifying the corresponding integer. If you'd like to have user-provided reference values, please implement this functionality and make a pull-request.

Why the name codonW?

Excerpted directly from John Peden's CodonW README...

Well first you must realise that "clustal" (a very popular multiple alignment program by Des Higgins) was originally written in Paul's lab in Trinity College Dublin. Clustal has since been rewritten from FORTRAN into C and undergone several name changes clustal-> clustalv-> clustalw -> clustalx. There was also a program called "codons" written in FORTRAN by Andrew Lloyd (a post-doc in Paul's lab), this was the original inspiration for codonW. An early version of codonW, written in C, was called codonv. When the code was enhanced to include multivariate analysis, what better name than codonW.