Assembly

Phables requires either short or long read sequencing data from metagenomic samples to be assembled. The following steps explain the steps required to be carried out beforehand.

Paired-end read files for short read assembly

Please make sure that the names of the read files are in the following format. Assuming that your paired-end sequencing reads are in the folder fastq, please make sure that the reads are in the format {sampleName}{pattern}{fileExtension}. fileExtension can be .fq, .fastq, .fq.gz or .fastq.gz.

Please make sure that your file pattern matches one of the following patterns.

_R1_ and _R2_
_R1. and _R2.
.R1. and .R2.
.R1_ and .R2_
_1_ and _2_
_1. and _2.
.1. and .2.
.1_ and .2_

For example, your read files can be

sample1_R1.fastq.gz
sample1_R2.fastq.gz
sample2_R1.fastq.gz
sample2_R2.fastq.gz
...

or

sample1_1.fq.gz
sample1_2.fq.gz
sample2_1.fq.gz
sample2_2.fq.gz
...

Long read assemblies

If you are using long read datasets, there is no specific naming format for the read files.

Assemble the samples

Phables requires the assembly graph file in Graphical Fragment Assembly (GFA) format. You can use any assembler that produces the assembly graph in GFA format to assemble your samples OR you can convert a FASTG file to GFA format.

If you have multiple samples you can pool together reads and do a co-assembly.

Recommended assemblers and tools

MEGAHIT

You can use MEGAHIT to assemble your paired-end short read data.

megahit -1 reads_1.fastq -2 reads_2.fastq -o megahit_out

By default, MEGAHIT does not produce an assembly graph file. You have to to run contig2fastg command from the MEGAHIT toolkit to build the assembly graph file. contig2fastg requires you to input the k-mer size used for the assembly. You can get the k-mer size from the contig IDs in the final.contigs.fa file. For example, you can use the grep command to print out the contig IDs as follows.

grep "^>" final.contigs.fa

Imagine you get the output as follows. Here the k-mer size is 141 as denoted by k141.

>k141_1456397 flag=0 multi=11.7570 len=1137
>k141_1235266 flag=0 multi=13.6963 len=1254
>k141_131192 flag=1 multi=47.8430 len=1510
>k141_1566081 flag=0 multi=9.6645 len=1372
...

Using the k value as 141, now you can run the contig2fastg command as follows.

megahit_toolkit contig2fastg 141 final.contigs.fa > final.graph.fastg

The MEGAHIT toolkit will result in a FASTG file which you can convert to GFA using fastg2gfa.

fastg2gfa final.graph.fastg > final.graph.gfa

If you want to run Phables on an assembly from a different k value found in the MEGAHIT output folder intermediate_contigs, please make sure to build the .fastg file from the .fa file with the corresponding k value. For example, if you want to run Phables on the contigs from k99.contigs.fa, you should first build the corresponding k99.graph.fastg file and then run fastg2gfa as follows.

megahit_toolkit contig2fastg 99 k99.contigs.fa > k99.graph.fastg
fastg2gfa k99.graph.fastg > k99.graph.gfa

metaSPAdes

You can use metaSPAdes to assemble your paired-end short read data.

spades.py --meta -1 reads_1.fastq -2 reads_2.fastq -o metaspades_output -t 16

After the assembly finished, the output will contain the assembly graph file as assembly_graph_after_simplification.gfa.

metaFlye

You can use metaFlye to assemble your long read data.

flye --meta --nano-raw reads.fasta --out-dir metaflye_output --threads 16

After the assembly finished, the output will contain the assembly graph file as assembly_graph.gfa.

Hecatomb

You can use Hecatomb which is a viral analysis pipeline to obtain a pooled assembly of your short read or long read data contained in a folder named reads. You can run hecatomb as follows. Note that you only need to run the assembly module to process your data for Phables.

hecatomb run --reads reads/ assembly 

After the assembly finished, the output will contain the assembly graph file as cross_assembly.gfa.

Now we are ready to run Phables.