About contigs taxonomy

User Questions
1

Hi ~

I have some questions when using the Megan.

  1. In the Megan published articles, it may be more suitable for reads taxonomy. But some articles that quoted Megan adopted different ways. Some articles predicted ORFs of contigs or MAGs firstly. Then blast ORFs to database.Finally get taxonomy of contigs or MAGs(https://doi.org/10.1016/j.fm.2021.103762;https://doi.org/10.1038/s41396-021-00932-y).Whether megan suitable to get the taxonomy of contigs use this method?
  2. I want to get viral contigs taxonomy using Megan. I followed such methods. And I can get the ORFs taxonomy. But one contigs may have several ORFs. How can I get each contigs taxonomy from ORFs informations? I know megan has gene centric assembly methods, but I want to using assembly contigs for other analysis, so I can only take conventional assembly measures such as spades. So how can I get each contigs taxonomy when using spades assembly ?
  3. About the unclassify definition I am in a puzzle. Maybe some annotation level only get family level.Some annotation level can get species level but have unclassify family or genus.Should I definte this two situation as unclassify ?

Looking forward to get professional advice. These problems have bothered me for a long time.Thank you very much.

2

To 1: yes, you can also use MEGAN to display the binning of ORFs or contigs.
When analyzing contigs, use the DIAMOND and MEGAN long-read modes.
To 2: you can look at the taxonomic assignment of ORFs along a contig using MEGAN’s “long read inspector”, in that dialog, set the “Layout” to Taxonomy.

image
image2836×1784 1.26 MB

To 3: The Unclassified node in MEGAN contains those sequences that have alignments, but for which MEGAN couldn’t figure out the taxon (for various reasons, e.g. the taxonomic identity of the reference sequences is not known to MEGAN).