Question: Metagenomics community profiling approaches.
2
gravatar for f.a.galkin
13 months ago by
f.a.galkin40
Germany
f.a.galkin40 wrote:

After I map a metagenome against a DB, I don't get an abundance table per se. I need to somehow decide, how I treat non-unique mappings and normalise data.

Centrifuge suggests an EM-based method, but what are the alternatives? I can try basic RPKM that completely ignores the problem, or I can try to assign each read to one of its taxons randomly. Any other options?

ADD COMMENTlink written 13 months ago by f.a.galkin40
1

Hi there,

Could you please elaborate what you mean by

After I map a metagenome against a DB

which tools, databases and methods are being used at this stage? Is it read-based or kmer-based mapping?

ADD REPLYlink modified 13 months ago • written 13 months ago by Sej Modha4.4k

K-mer based, Centrifuge, mapped onto their index of bacterial/archaeal genomes

ADD REPLYlink written 13 months ago by f.a.galkin40

Doesn't centrifuge output file centrifuge_report.tsv contain a measure of abundance? Did you look at the output from centrifuge-kreport command?

ADD REPLYlink modified 13 months ago • written 13 months ago by Sej Modha4.4k

It does contain, but I am looking for alternatives, if there are any. Centrifuge's report assigns zero abundances too generously

ADD REPLYlink written 13 months ago by f.a.galkin40

I do not have an answer, but should you not only use certain markers to find the taxonomy? Markers like CO1, 16S, 18S, ITS etc. Because you say you the whole metagenome. There are genes that are in almost every organism so just take a random taxonomy would not be good.

ADD REPLYlink modified 13 months ago • written 13 months ago by gb880

Marker gene based method are biased and do not guarrantee detecting all the present organisms

ADD REPLYlink written 13 months ago by f.a.galkin40

"Marker gene based method are biased" True, if you do amplicon sequencing and a PCR. But you have a metagenome as I understand from your question. You can not just use any gene to find the taxonomy. Maybe metaphlan can help: http://huttenhower.sph.harvard.edu/metaphlan

ADD REPLYlink written 13 months ago by gb880
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 565 users visited in the last hour