Input assembled metagenoms

[lalalagartija]

Hi, Can I use Kraken for the classification of my assembled data ? Are there special options for that ?

Thank you very much in advance,

[MichaelAngelou]

Hi @lalalagartija I just want to ask if your'e able to run kraken2 using assembled metagenomes as an input? Thank you very much.

[lalalagartija]

Hi, actually yes it worked but the problem is that the sequence use to match to a toxonomy on each contig is very short so I wouldn't suggest this method. I finally ran Kraken on reads instead.

[MichaelAngelou]

Hello @lalalagartija thank you very much for the reply. How long (bp) is your contigs? Mine is >500bp. So you used the clean paired end reads for the classification?

[lalalagartija]

Mine were over 5000pb. I used the reads for classification with Kraken and the assembled contigs for the classification via blastp (diamond). What would you need to do ?

[MichaelAngelou]

I basically just want to determine the diversity of microbial taxa across diff samples, so i) classify the assembled contigs from metaspades using kraken, and 2) determine their relative abundances (using braken). Then calculate the alpha/betadiversity. i read many papers stating that kraken performs better than other metagenome classification tools.

[lalalagartija]

I don't know braken but I also fould that kraken outperforms the other tools. I don't know if you win anything from assembling the contigs prior to kraken because the match used to classify are very short. Although, the contigs can be usefull for further analysis.

[MaryoHg]

Hi, @lalalagartija @MichaelAngelou

Just to say that AFAIK, Bracken y Kraken were supposed to work together. Bracken 'improves' Kraken classification. Check it out.

May I ask something: Did you use the quality filtered reads for kraken and discard the singles (unpaired sequences with quality filtered passed) or did you cat those?

Thanks.

[isardi]

Hi @MichaelAngelou, I have been trying to figure out how to calculate alpha and beta diversity using the Bracken output with no luck. What package do you use? Did you have to modify the output in anyway? I spent the whole day trying to figure this out so your input is very much appreciated. Thanks in advanced!

[ciemanek]

Hi @isardi !

You can use R package vegan for calculating the diversity indices on Bracken output :)

[herokoking]

Hi @lalalagartija ： I have the same idea with you about this option ： --- “the assembled contigs for the classification via blastp (diamond)“ Could you give out some demo code for this purpose. THX

[lalalagartija]

@MichaelAngelou I also wanted to do alpha and beta diversity but it turns out that alpha diversity is quite codified for species level/OTU so you can hardly do it with kraken output but for beta diversity it is just fine ! There must be some ways of transforming he data for alpha diversity but I don't know how. For beta diversity you have to transform your data in an OTU table like. What does the output looks like ?

@herokoking I would be happy to share my code but I didn't wrote it. I use one from a collaborator but I will ask him to publish it. In the mid time you can check "Lowest Common Ancestor" at 50% cutoff. That is what I used. You apply that on the output of a diamond blast against refseq or nr (or whatever database) asking for "taxid" and from this taxid you ask for the lineage either going through entrez but it is very long. I use a modified krona script that is not mine neither. I can ask to publish that too if you don't find it elsewhere.