whether Taxonomy barplot made in R from squeezemeta output represents taxonomy of reads or assembly

[mohitsharma-123]

Hello
I want to know that i have ran squeezemeta on my 4 samples using metaspades and coassembly mode.
I m using R workflow provided in this program for analyzing my squeezemeta output.
I want to know whether the taxonomy plots generated will represent taxonomy at reads level or of assembly.
Also when i m using tools like kaiju, centrifuger on both reads from fastp and metaspades assembly i found huge difference in taxonomy and read taxonomy is aligning more with squeezemeta output

plotTaxonomy(Hadza, rank='phylum', count='percent')

[jtamames]

Hello
If you did co-assembly, the all the information refers to the coassembly. You won´t have any classification of reads unless you use sqm_reads.pl or sqm_longreads.pl
It is not rare to have differences between these other tools and squeezemeta since they use different methods for the annotation. Check our paper on this subject if https://doi.org/10.1186/s12864-019-6289-6

Best,
J

[mohitsharma-123]

So, for metaspades assembly taxonomy which one is more preferable Squeezemeta, kaiju, kraken2, centrifuger
also for reads level which one you suggest.
Thank you very much for your reply really appreciate your response.
Thanks for wonderful program.

[jtamames]

SqueezeMeta, obviously! lol