add canu assembler

close #12

README.md

@@ -9,7 +9,7 @@ See the preprint here: [Snakemake Workflows for Long-read Bacterial Genome Assem
 
 | read filtering | assembly | long read polishing | short read polishing | reference-based polishing |
 | --- | --- | --- | --- | --- |
-| [Filtlong](https://github.com/rrwick/Filtlong) | [Flye](https://github.com/fenderglass/Flye)<br/> [raven](https://github.com/lbcb-sci/raven)<br/> [miniasm](https://github.com/lh3/miniasm)<br/> [Unicycler](https://github.com/rrwick/Unicycler) | [racon](https://github.com/lbcb-sci/racon)<br/> [medaka](https://github.com/nanoporetech/medaka) | [pilon](https://github.com/broadinstitute/pilon/wiki)<br/> [Polypolish](https://github.com/rrwick/Polypolish)<br/> [POLCA](https://github.com/alekseyzimin/masurca#polca) | [Homopolish](https://github.com/ythuang0522/homopolish)<br/> [proovframe](https://github.com/thackl/proovframe) | 
+| [Filtlong](https://github.com/rrwick/Filtlong) | [Flye](https://github.com/fenderglass/Flye)<br/> [raven](https://github.com/lbcb-sci/raven)<br/> [miniasm](https://github.com/lh3/miniasm)<br/> [Unicycler](https://github.com/rrwick/Unicycler)<br/> [Canu](https://github.com/marbl/canu)  | [racon](https://github.com/lbcb-sci/racon)<br/> [medaka](https://github.com/nanoporetech/medaka) | [pilon](https://github.com/broadinstitute/pilon/wiki)<br/> [Polypolish](https://github.com/rrwick/Polypolish)<br/> [POLCA](https://github.com/alekseyzimin/masurca#polca) | [Homopolish](https://github.com/ythuang0522/homopolish)<br/> [proovframe](https://github.com/thackl/proovframe) | 
 
 
 ## Quick start
@@ -153,6 +153,11 @@ Default assembly. Miniasm does not do any polishing by itself.
 **`unicycler`**  
 Unicycler does a hybrid assembly, i.e., both ONT and Illumina reads must be present in `fastq-ont` and `fastq-illumina`, respectively.
 
+### Canu
+
+**`canu`**  
+The Canu assembler requires to know the genome size (in Megabases) beforehand, use Snakemake option: `--config genome_size=5.2` (e.g. for 5.2 Mb)
+
 ### racon
 Following keywords can be used to polish an assembly using ONT reads:
 

Snakefile

@@ -26,6 +26,12 @@ if config.get("map_medaka_model", False):
     table = pd.read_csv(medaka_model_file, sep="\t", lineterminator="\n", header=None)
     map_medaka_model = dict(zip(table[0], table[1]))
 
+# this option is needed by Canu
+target_genome_size = None
+if config.get("genome_size", False):
+    target_genome_size = config["genome_size"]
+    print("Target genome size = " + str(target_genome_size) + "M")
+
 
 wildcard_constraints:
     sample="[^_]+",
@@ -74,6 +80,14 @@ if not references_protein and [
         "Error: must provide at least one reference protein file when using proovframe"
     )
 
+# if any desired assembly uses Canu then need to provide target genome size
+if not target_genome_size and [
+    string for string in sample_assemblies if "_canu" in string
+]:
+    quit(
+        "Error: must provide target genome size when using Canu assembler, use option (e.g. for 5.2Mb): --config genome_size=5.2"
+    )
+
 list_outputs = expand(
     "assemblies/{sample_assembly}/output.fa", sample_assembly=sample_assemblies
 )
@@ -372,6 +386,25 @@ rule ravenX:
         """
 
 
+rule canu:
+    conda:
+        "env/conda-canu.yaml"
+    threads: 10
+    input:
+        fqont=get_ont_fq,
+    output:
+        fa="assemblies/{sample}_canu/output.fa",
+        link="assemblies/{sample}_canu.fa",
+    log:
+        "assemblies/{sample}_canu/log.txt",
+    shell:
+        """
+        canu -nanopore -d assemblies/{wildcards.sample}_canu/ -p output useGrid=false maxThreads={threads} genomeSize={target_genome_size}m {input.fqont} >>{log} 2>&1
+        cp assemblies/{wildcards.sample}_canu/output.contigs.fasta {output.fa} 2>>{log}
+        ln -sr {output.fa} {output.link}
+        """
+
+
 # for running racon once
 rule racon:
     conda:

env/conda-canu.yaml

@@ -0,0 +1,9 @@
+name: ont-assembly-snake-canu
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+  - anaconda
+dependencies:
+  - bioconda::canu=2.2
+