Adding Unicycler to the workflow

[danejo3]

The purpose of this PR is integrate Unicycler.

Unicycler is known for its hybrid assembly pipeline for bacterial genomes. When dealing with short reads, unicycler will use spades and optimize it to:

• Try a wide range of k-mer sizes and automatically selects the best.
• Filter out low-depth parts of the assembly to remove contamination.
• Apply SPAdes repeat resolution to the graph (as opposed to disconnected contigs in a FASTA file).
• Reject low-confidence repeat resolution to reduce the rate of misassembly.
• Trim off graph overlaps so sequences aren't repeated where contigs join.

This PR will resolve #11 .

[danejo3]

As of right now, YEAT can only process illumina paired-end reads. In the future, I plan on adding additional support for unpaired reads and long reads.

Unicycler has support for all kinds of reads--paired-end, unpaired, and long. One thing that is unique to unicycler is its ability to do a hybrid assembly with both short and long reads. According to the unicycler documentation, the hybrid assembly will be the most accurate (which makes sense).

unicycler -1 {read1} -2 {read2} -l {long_read}

In this PR, I have provided support for unicycler but only paired-end reads. In another PR, we'll need to add the support for all kinds of reads. In consequence, we may need to refactor a large part of the code base to enable the other read types.

[danejo3]

Another thing that is unique with unicycler is that it creates its own Bandage file called assembly.gbk.

YEAT does not support bandage yet.

It looks like spades creates a few file(s) for bandage. -assembly_graph.fastg -assembly_graph_after_simplification.gfa -assembly_graph_with_scaffolds.gfa

megahit -> megahit_toolkit -> contig2fastg https://github.com/voutcn/megahit/wiki/Visualizing-MEGAHIT's-contig-graph

[danejo3]

Unicycler has some interesting options: --min_fasta_length (default 100) --mode (conservative/normal/bold)