Is there any way to fix the low percentage mapping results ?

Question

[ssko]

Hi Guys!

Is there any way to fix the low percentage mapping results ?

I have some shotgun metagenome data. Before mapping I checked the quality with FAST QC and then trimmed it. Minimum contig length was fixed at 200 bp. Then, my samples were mapped with Bowtie2. Below, there are the some mapping results:

# Sample    Total reads Mapped reads    Mapping perc    Total bases
#samples13  21326256    17067391    80.03           1615843043
#samples14  109567          15568           14.21             7605088
#samples15  258309          92459           35.79             18786001
#samples16  339971          91735           26.98             24911153

[Michael]

You are aligning your metagenome assembly to what exactly? I think, if you are doing alignments on the nucleotide level using bowtie you are not going to get good cross species alignments. Therefore, use a more sensitive aligner e.g. blast, or align the translated sequence, e.g. blastx, and what would help most if you add more organisms that really are there or close relatives to your database. I think that bowtie is the wrong tool for metagenomics in most cases, unless you now the organisms exactly, but then what is the point of a metagenomics approach? Even then, there are more sensitive tools than bowtie.