Closed thierryjanssens closed 4 years ago
Update: Currently the mgkit
method is being validated. We expect to incorporate it in the next version.
@thierryjanssens , do you have an update on this issue? Does the mgkit
LCA implementation in the v0.9.5-dev
branch fix this?
We're currently using Mgkit as a default instead of krona. considering this as fixed
When analysing my dataset with the PZN_ERVINGS fork of PZN (the name of the current pipeline before it was renamed to Jovian), that contains a.o. an alternative LCA method (mgkit; Rubino, F. and Creevey, C.J. 2014. MGkit: Metagenomic Framework For The Study Of Microbial Communities. . Available at: figshare [doi:10.6084/m9.figshare.1269288].`), I observed that Jovian is missing certain viral contigs. The minimum contig length was reduced to 250 bp, in order to increase the detection in the dilute respiratory samples...
e.g. a Coronavirus OC43 and Adenovirus.
Despite the good blast hits (with decent e-values and bitscores) Krona place the deduced taxonomic position in the root of life. e.g for the adenovirus case:
I will work on an implementation of the mgkit tools in Jovian ASAP.