I have been able to use the Kraken2 pipeline for analysing the nanopore reads and derive a KronaPlot for metagenomics. Since KronaPlot also gives a list of reads from a sample that is associated with the species identified, I am able to find those reads from my sample fasta file. Now, when I nucleotide BLAST those reads in NCBI, they don't identify the same species that the Kraken2 pipeline has output. Also, if I try to assemble the reads to the species identified by Kraken2, they don't align.
What am I doing wrong here, or did I miss identifying some important concept?
I am using Standard Kraken2db (All bacteria, fungal, viral and human) available here
https://benlangmead.github.io/aws-indexes/k2
I dont think you are doing anything wrong. Different software/algorithm and database can create big differences in identifying taxa.
Article The selection of software and database for metagenomics sequ...
In addition to Péter Gyarmati comment on software, you should be aware that what method is used to extract can also causes differences in your downstream analysis. Here is one such example, but there are many others if you search "metagenomic differences based on extraction methods".
Article Choice of bacterial DNA extraction method from fecal materia...
Thanks for the input. I think I know what is the problem. This is discussed at length in Github which I failed to find earlier. I need tto probably redo the code and assign Confidence Cut-off
https://github.com/DerrickWood/kraken2/issues/265
The two software uses different algorithms, it is not surprise that you got different result. Nanopore data often have high level of error rate, while Blast was initially designed for accurate alignment. That might be a reason. In addition, adjusting the confidence value of Kraken2 can result in large variability of sequences which can be taxonomically annotated.
Thank you for the input. I found that redoing the analysis with confidence scores in Kraken and then further analysis with Bracken brought in some consensus and better confidence in the final output.
- Badges
- Science topic
- Similar topics
- Biological Science
- Anatomy
- Tissue