Efficient genome assembly is a growing need due to the introduction of high volume sequencing such as amplicon and shotgun sequencing. Novel ways to quickly assemble genomes and reduce error margins have become an important commodity. K-mers are a popular tool, allowing for high error margins to be reduced significantly by parsing the read and matching together sequences. KREPE is a python command-line tool created to better serve the bioinformatics community in creation of meaningful graphs and calculations. By utilising python dictionaries and multiple different graph visualizations, KREPE serves as a one-stop shop for assembling genomes, creating occurrence graphs, assembling genomes with De Bruijn graphs, calculating Jaccard similarity, generating venn diagrams, and making high quality dendrograms. KREPE is an open-source tool for analyzing and creating representative genome visualizations.
Co-Authors Erika Pedersen and Aengus McGuinness created KREPE at a fellowship with the Institute for Computing in Research.
In bioinformatics, it's important to be able to use strands of DNA to compare organisms, understand and assemble genomes, and get a better picture of how everything works together in an organism. A common tool is the k-mer (nucleotide sequences of length k). When a biologist is working on figuring out a genome, or comparing a cat and a dog, they start with the k-mers. The way they work is similar to a paper shredder; you put in a paper (or multiple reads of the same sequence), and shred it into bits. Then, you take all the paper bits and piece them back together in the same way as before, by finding similarities in each of the pieces. This is useful to biologists because sometimes there are errors when the machine reads a sequence, so with multiple copies it's simple to find errors and toss them out early so they don't affect the overall 'picture'.
These k-mers are sometimes strung together to make stringy looking graphs called 'De Bruijn Graphs', that are used to find a pathway through a genome- this process of finding a pathway is called collapsing. 'Collapsing' a De Bruijn graph is comparable to cobbling together different bus or train routes that start at the place where the last one ends, and finding a way to your destination.
Another kind of graph that is useful is a bar graph. Bar graphs show the appearances of k-mers and can be used to determine the quality of reads and determine what type of organism is being seen.
Dendrograms can be used to compare any number of organisms to each other and see how genetically far apart they are. They have a vertical line from the bottom left to the top right that demonstrates samples intersecting with themselves, and the rest of the squares show intersections between different samples. Venn diagrams are another way to demonstrate the overlap of organisms.
Our program calculates Jaccard similarity (which is basically the amount that things have overlapping, 1.00 being completely the same and 0.00 being completely dissimilar) and uses it for the dendrograms and the venn diagrams.
The goal of the project is to create a streamlined, easy to use package and command line tool that allows users to count k-mers and easily generates the following:
- A list of the occurring k-mers in the given file that occur more than once
- A bar graph that shows the frequency of said k-mer occurrence
- A De Bruijn graph that can then be used for genome assembly
- A dendrogram that models genetic similarity and metadata
- A Venn Diagram that models genetic similarity
The package is protected by copyright under the GNU General Public License
Use the package manager [pip] (https://pip.pypa.io/en/stable/) to install krepe
$ pip3 install krepeOR you can install without pip by using git:
$ git clone https://github.com/RGBwalnut/KREPEThe inputs that are allowed are .fasta, .fastq, and .fna.
This package is completely open-sourced and is being updated by authors Erika Pedersen and Aengus McGuinness as needs arise
Contributions will be accepted, in the instance that the contributor is willing to conform with the following:
A. be listed as a minor contributor in the description
B. not be listed under the copyright
C. be willing to hear suggestions on minor changes to the proposed contribution
As well as in the instance the authors deem the contribution useful.