wernerkrampl/perfect-blue

CovEst

Tool that estimates coverage (and genome size) of dna sequence from reads.

Important

Prediction models have been renamed in version 0.5.9 to match names in publications.

Changes are as follows:

Basic -> E

Repeats -> RE

Basic_Polymorphism -> EP

Repeats_Polymorphism_Equal -> ERNPE

Requirements

• python 3.4+
• python3-dev
• gcc

Installation:

pip install -e . from the project directory

Usage

type covest --help for the usage.

Basic Usage:

covest histogram -m model -k K -r read_length

• You can specify the read file using -s reads.fa parameter for more precise genome size computation.
• default K is 21
• default read length is 100
• currently, the supported models are:
  • basic: for simple genomes without repeats
  • repeat: for genomes with repetitive sequences

Input Histogram Specification:

The input histogram can be generated from the read data using jellyfish.

• jellyfish count -m K -C reads.fa -o reads_table.jf -s 1000000000
• jellyfish histo table.jf -o reads.hist

The format of the histogram is just list of lines. Each lines contains an index and value separated by space.

Output Specification:

CovEst outputs it's results in simple subset of YAML format for best human readability and possibility of machine processing.

The output are lines containing key: value. The most important keys are coverage and genome_size (or genome_size_reads if reads size was specified).

Other included tools

• geset.py tool for estimation genome size from reads size and known coverage
• reads_size.py tool for computation of the total reads size
• kmer_hist.py custom khmer histogram computation, it is much slower than other tools, so use it only if you have no other option.
• read_sampler.py script for subsampling reads, useful if you have very high coverage data and want to make it smaller.
• fasta_length.py get total length of all sequences in fasta file.

Copyright and citation

This section is applicable to original CovEst (0.5.6) by M. Hozza.

Original CovEst is licenced under GNU GPLv3 license.

CovEst is research software, so you should cite us when you use it in scientific publications!

Hozza, M., Vinař, T., & Brejová, B. (2015, September). How Big is that Genome? Estimating Genome Size and Coverage from k-mer Abundance Spectra. In String Processing and Information Retrieval (pp. 199-209). Springer International Publishing.