Hello all,
I have assembled(using velvet) illumina contigs for taxonomic classification. Is there any tool which can do the taxonomic classification of assembled illumina reads.
I tried MEGAN. But MEGAN is only considering the read counts. I also want to take the coverage information of each contig in to consideration.
Please give me some advise here.
Thanks, Deepthi
•
written
by
deepthithomaskannan
1 answer
The MEGAN paper describes the method that they use to assign taxonomical ranks.
If I remember it correctly it is a very simplistic majority win, perhaps weighted by alignment length.
I would say that the approach is more suitable for short read alignments where there aren't many competing options. For long contigs it is hard to accept that their method is anything but a rough approximation.
But in lieu of better option is as good as it gets.
Log in to answer this question.
-
How to extend assembled PACBIO contigs for scaffolding
written by K •
-
to calculate coverage of each contig after assembly
written by singh.jyotika •
-
Tool For Taxonomy Classification Of contigs and reads
written by blancai.taboada •
-
Velvet: What's the relation between kmer coverage and normal coverage?
written by novice
-
Inferring most probable integer copy number (repeat count) of contigs of a draft genome
written by misaghb •
-
Copy number (repeat count) of assembled contigs
written by misaghb •
-
MEGAN and Kraken for taxonomy classification
written by deepthithomaskannan
-
MEGAN or Karaken
written by deepthithomaskannan
-
Total Contig Length From Velvet Output
written by deepthithomaskannan
-
Good Illumina Assemler
written by deepthithomaskannan
doesn't the MEGAN tool use the output of a Blast to determine the taxonomical rank? What I am saying is that MEGAN is not a classifier it only processes alignment outputs that you create with other tools
You are right. I gave blastn(contig file against nt DB) file as the input of MEGAN. MEGAN is giving me some output. Since the reads are assembled contigs the weight of each contig will be different according to the coverage. Hence the output of MEGAN will not be correct. I want to do oraganism detection and % of each organism etc using metagenomic illumina sequences.
Why not just use raw reads then?
I tried raw reads. One: It is taking so much time to complete blastn. Two: Results are no better. Three. blastn files are so big. I am trying with other tools other than MEGAN.