handibles
Jamie.FitzGerald -a- MTU.ie
:: Species-level 16S databases for Kraken2:
Big fan of Kraken2(+Bracken!), but the existing Silva 138 database creates issues with assignment to species level (Silva gives each species the taxid from it’s parent genus - so Kraken2 can only go as far as genus). GreenGenes2 requires an even larger degree of wriggling in order to parse the database.
Here we work around those issues for anyone working on 16S assignment:
GreenGenes2 16S species-level database for Kraken2:
code& walkthrough: approach and code for for making species-level Kraken2 databases (v2024.09).- download subset K2 db: 1.3-1.65kbp subset (331K seqs, 170MB)
- download complete K2 db: complete (21M seqs, 1.3GB)
Silva 16S species-level database for Kraken2:
codeonly: sparse code for making species-level Kraken2 databases (v138.2).- download complete K2 db:
NB: both Silva and Greengenes2 use taxids that do not match NCBI taxids. The taxonomies (names!) are fine, and if you are not using taxids, this should not matter to you at all. For Silva 138.2, we assign new (fake) taxids to all ranks below genus (i.e. species, strains, subspecies, submarines, substrains). With Greengenes, this code assigns completely new (fake) taxids to all ranks.
See also a list of other possible issues with this approach, and feel free to list your own, like my concern about reference length…
:: CLI, Microbial Ecology, and R - CLIMBER guide to Microbial Ecology
Some basic steps in microbial ecology, focusing on the processing of 2ndGen Illumina fastq data, into either amplicon (e.g. 16S) or metagenomic (e.g. shotgun) datasets, followed by ecology-based analysis of the communities and patterns we find in that data. Now with less emojis.
check out
climber
:: Own work
You’d think this would be the first thing to go in, but no.
:: things to finish writing up
the Centre Log-Ratio transform:
- CLR transform and pathways (many one-to-many relationships)
- ALR transform and pathways (a single set of one-to-many relationships)
- effect of Multiplicative Replacement (
zCompositions::cMultRepl) in the CLR transform - effect of sequencing depth on residuals of linear modelling of CLR values
- effect of sequencing depth on residuals - employ Lasso on CLR values?
- subsetting ALDEx2 datasets (…really don’t remember doing this)
Microbial Diversity:
- all the promised
Recology stuff - how to choose 40 colours and live
- subsetting databases to sequence length
Ecological Diversity:
- personal walkthrough of Willis’ class stuff on rethinking alpha diversity
- Bray-Curtis on transformed data
- Alpha diversity omnibus testing (KW, F, W)
- Dirichlet Multinomial Modelling (DMM) and diversity
- how much does Bracken affect ecological diversity? (not a huge amount)
- toy diversity in data
- filtering features: k & A
- singleton effects in diversity
CLI:
- downloading files from RefSeq/Gene/NCBI/
entrez-direct - downloading files from ENA
- downloading files from illumina
- renaming accessions for KMA/CCMetagen
- hashing over dictionaries in bash for building databases (KMA/CCMetagen)
- PLS-DA-Kfold.R
#thing4
It’s not any sheer to put underneath with an unless - see with and folder. But, did and with the place of not yet although with and also? - making presentation un up