So, you have generated an assembly of your metagenome from short-read data. What now? How can you tell how good it is?
Now, onto getting quantitative metrics:
Now we can run a few stats on our assembly. To do this we will use QUAST:
cd ~/
git clone https://github.com/ablab/quast.git -b release_4.5
export PYTHONPATH=$(pwd)/quast/libs/
Now, run QUAST on the assembly:
cd ~/assembly
mkdir quast-evaluation
cd quast-evaluation
ln -fs ../combined/final.contigs.fa megahit.contigs.fa
~/quast/quast.py megahit.contigs.fa -o megahit-report
cat combined-report/report.txt
What does this say about our assembly? What do the stats not tell us?
For thought and discussion: a few nice slides from Jessica Blanton on assessing assembly quality.
The stats that are reported by QUAST do not mean much on their own– for them to be more meaningful it is helpful to have another assembly against which to compare them. For that reason we have assembled the same data using a different assembler– metaSPAdes. To see how the assembly was generated and to try running the assembly on your own later you can follow the metaSPAdes tutorial. It took ~2 times as long as MEGAHIT on this dataset.
So, for now, download the metaSPAdes assembly:
cd ~/assembly/quast-evaluation/
curl -LO https://osf.io/h29jk/download
mv download metaspades.contigs.fa.gz
gunzip metaspades.contigs.fa.gz
Now, adjust the scripts used previously to calculate the same metrics for the new assembly. How do the two compare? What metrics should you care about?
Let’s look at the two reports in parallel!
paste */report.txt | cut -f1-2, 4
How do they compare? What does it mean? Does quast give us enough information to tell which assembly is better?