Using Kraken2 with the human telomere to telomere genome (T2T)

[tatianaorli]

What do you mean by "agnostic"?

The mapper bowtie2 was not developed specifically for mapping bacterial genomes or metagenomics. It is a generic tool for mapping. Some tools are more tailored to metagenomics mapping. They take into account the differences in coverage between the species due to abundance and know how to differentiate between two close species that have different coverage overall.

[konstantia_lynch]

My question is referring only to Kraken2, and the package offers its own databases (build in the past by the developers), and Kraken also offers the option to build a personalized database. I used recently bowtie2 with T2T and it was very straightforward, no special commands were needed for a good mapping of my data (human reads had high quality). I think that you could check other metagenomics-focused mappers, because bowtie is more of an agnostic tool.

What do you mean by "agnostic"?

[tatianaorli]

My question is referring only to Kraken2, and the package offers its own databases (build in the past by the developers), and Kraken also offers the option to build a personalized database. I used recently bowtie2 with T2T and it was very straightforward, no special commands were needed for a good mapping of my data (human reads had high quality). I think that you could check other metagenomics-focused mappers, because bowtie is more of an agnostic tool.

[konstantia_lynch]

Hello tatianaorli. I have just started working with T2T assembly instead of hg38. Currently, I am using Bowtie2 to remove human DNA from my soil shotgun metagenomic datasets. I am following the default Bowtie2 guidelines and have built the indexed database using the 'bowtie2-build' command. The T2T assembly was downloaded from this link: https://www.ncbi.nlm.nih.gov/dataset...A_009914755.4/

I also should state that I am a beginner in metagenomics; came here to check if someone can advise against my way of performing this step of analysis.

[tatianaorli]

Has anyone tried and succeeded in using the newest human genome version (T2T) in Kraken2 instead of the human hg38?

There are rumors that it can improve read assignment when dealing with host/microbiome mixed metagenomics. In my case, the microbial data is low abundance, and I need to eliminate the maximum amount of human reads and prevent false positives (to the microbial hits).
Which commands did you use? How did you build the reference?

Thanks!

Update: I ended up creating a custom database that includes T2T-CMH13 as the human reference and other genomes, and used my original parameters with it.