What is the resolution of microbial sequence assembly

Question

[marongiu.luigi]

Hello

I have assembled reads with SPAdes and obtained a list of bacterial species such as:

Prevotella intermedia
Prevotella intermedia ATCC 25611 = DSM 20706
Prevotella scopos JCM 17725
uncultured bacterium
uncultured bacterium 3e18
Enterobacter hormaechei subsp. oharae
Fusobacterium hwasookii ChDC F174
Fusobacterium nucleatum subsp. animalis
Fusobacterium nucleatum subsp. animalis 4_8
Fusobacterium nucleatum subsp. animalis 7_1
Fusobacterium nucleatum subsp. nucleatum ATCC 25586
Fusobacterium nucleatum subsp. polymorphum
Fusobacterium nucleatum subsp. vincentii 3_1_27
Fusobacterium nucleatum subsp. vincentii ChDC F8

These species are from the same sample. My question is: shall I trust the subdivision at subspecies (or isolate)? Or shall I cut at the species level? In that case, I would get:

Prevotella intermedia
Prevotella scopos
uncultured bacterium
Enterobacter hormaechei 
Fusobacterium hwasookii 
Fusobacterium nucleatum

In the first case, I'll get 14 hits, in the second 6. Which one is more trustworthy?

Thanks

[GenoMax]

How did you obtain this list? By blast searching your contigs?

[marongiu.luigi]

yes, that is right

[5heikki]

And what kind of criteria are you using? If a 500,000 bp contig gets a significant 500 bp alignment against something, does it mean anything at all (other than possible HGT perhaps)?

[juanjo75es]

It depends on your algorithm for selecting contigs and selecting blast results. If you just choose the first match from short contigs then it's likely that some subspecies matched are just casual and you should consider the second list. No way to know without more information.