Low value of mapped reads against a metagenomic co-assembly

Question

[Biok]

Hi, everyone, I have two metagenomic samples, I co-assembled these samples together using both spades and Megahit and then I mapped each sample against the co-assembly using bowtie2. I got for each sample around 40 % overall alignment rate. I would like to know if this low value is normal (given that the assembly comes from reads of both samples)???

[h.mon]

By default, megahit have a contig length cut-off of 200 base pairs, so a lot of reads may have been left out of the final assembly.

You can change the minimal contig length with the --min-contig-len command-line option:

 --min-contig-len  <int>  minimum length of contigs to output, default 200

[Biok]

thanks for the reply, I selected contigs larger than 1000 bp because my goal is to reconstruct Metagenome-assembled genomes (MAG's). Initially I got around 7,000,000 contigs from Megahit but after selecting contigs larger than 1000 bps I only got an final assembly of around 455,000 contigs and I used this assembly to map the filtered reads from each sample and I got 40 % overall alignment rate. Many contigs were removed, maybe that's the reason that explains the overall alignment rate.

Answer 1

[biobiu]

Its not necessarily low...

1. If it's host metagenomes you might have the host reads. Try mapping to the host genome.

2. If it doesn't help much would try assembly separately to each sample.

3. Remember that the assembly quality is affected by the read length, sequencing depth and the diversity of the samples.

Good luck :)