A pipeline and reference protein sequence data for generating core-proteome alignment and computing core-proteome amino-acid identity (cpAAI) values from bacterial genomes of the Rhizobiaceae family, to assess their membership to an existing or a new genus.
These software/data are in support of this paper:
Nemanja Kuzmanović, Camilla Fagorzi, Alessio Mengoni, Florent Lassalle, George C diCenzo. 2021. Taxonomy of Rhizobiaceae revisited: proposal of a new framework for genus delimitation. Int J Syst Evol Microbiol 72(3):005243. [https://doi.org/10.1099/ijsem.0.005243]
- Python 3 (recommended version >= 3.8.5), with package:
- Bio (BioPython) (recommended version >= 1.78)
- NCBI Blast+ (recommended version >= 2.6.0+)
- Clustal Omega (recommended version >= 1.2.3)
or - MAFFT (recommended version >= v7.487) (preferred)
- R (recommended version >= 4.0.2), with package:
- phangorn
The dependencies required to run the pipeline script genome2cpAAI.py can be satisfied by building the conda environment specified by the environment.yml file.
For visualisation purposes, R and its package phangorn can/should be installed spearately.
The data folder contains reference marker protein sequence sets and their corresponding alignments for 170 non-recombining core protein markers conserved among a set of 97 Rhizobiaceae genomes, including all available genomes of type strains of species described in the Rhizobiaceae family (as of April 2021).
The script genome2cpAAI.py in the pipeline folder can be used to obtain homolous sequence to the reference marker set described above, or to any analogous dataset, for instance a core-proteome dataset defined for another prokaryotic taxon.
The script genome2cpAAI.py takes as input a list of complete genomic nucleotide sequence (contig) files from the query organisms (argument -q), a list of files containing the reference marker protein sequence sets (argument -p) and a path for the folder where the output will be written (argument -o). Optionally, a list of files containing the alignments of the reference marker protein sequence sets can be provided (argument -A). All sequence and alignment files should be in Fasta format.
This script will locate in the input genome the loci coding for the protein homologs of the reference marker proteins using tblastn. It will then extract these sequences, translate them, and align them with the reference sequence set using mafft (the aligner clustalo can be used alternatively if specified through the argument --aligner); if the reference alignments were provided, these are used as a alignment guide using mafft -add algorithm, allowing faster processing.
Finally, the resulting core protein alignments are concatenated into a single file.
Here is an example of how to use the script genome2cpAAI.py on generic data (i.e. with any reference core protein set, pre-aligned or not):
# generate list of query genome sequences (use of .fna extension is just indicative here, any nucleotde fasta file will do)
ls my_query_genomes/*.fna > testgenomelist
# generate list of reference marker protein sequences (use of .faa extension is just indicative here, any protein fasta file will do)
ls my_ref_protein_sequences/*.faa > my_ref_protein_sequences_list
mkdir genome2cpAAI_out/ tmp/
# run the pipeline (assuming you have 8 cores to support multi-threaded computation)
genome2cpAAI.py -q testgenomelist -p my_ref_protein_sequences_list \
-o genome2cpAAI_out --threads 8 --tmp_dir tmpand if you have already aligned your reference protein set:
# also generate list of reference marker protein sequences (use of .aln extension is just indicative here, any aligned protein fasta file will do)
ls my_ref_protein_alignments/*.aln > my_ref_protein_alignments_list
# run the pipeline (assuming you have 8 cores to support multi-threaded computation)
genome2cpAAI.py -q testgenomelist -p my_ref_protein_sequences_list \
-A my_ref_protein_alignments_list \
-o genome2cpAAI_out --threads 8 --tmp_dir tmpIf you wish to estimate the cpAAI of a query Rhizobiaceae genome against our reference set of 170 marker proteins from 97 reference strains as described in Kuzmanovic et al. (2021), please use the following commands.
We strongly recommend using the pre-aligned reference protein files as we cannot guarantee that the cpAAI values derived from a third-party alignment will be consistent with those described in our manuscript.
# generate list of reference marker protein sequence and alignment files
# lists of files for the 170 Rhizobiaceae core protein markers are already available in data/
ls whereyouputyourrepo/cpAAI_Rhizobiaceae/data/protein_sequences/*.faa > protein_sequences_list
ls whereyouputyourrepo/cpAAI_Rhizobiaceae/data/protein_alignments/*.aln > protein_alignments_list
mkdir genome2cpAAI_Rhizob_out/ tmp/
# run the pipeline
genome2cpAAI.py -q testgenomelist -p protein_sequences_list \
-A protein_alignments_list -o genome2cpAAI_Rhizob_out \
--threads 8 --tmp_dir tmpThen, the output concatenated alignment concatenated_marker_proteins.aln (located in the specified result folder, here genome2cpAAI_out/) can be used to compute a core-proteome tree using the phylogenetic program of your choice (task not included in this package) and to compute the cpAAI values between query and reference strains.
This can be done with the following R script:
library(phangorn)
nfconcatprotaln = "genome2cpAAI_out/concatenated_marker_proteins.aln"
concatprotaln = as.matrix(read.FASTA(nfconcatprotaln, type='AA'))
cpaai = 100 * (1 - as.matrix(dist.aa(concatprotaln, scaled=TRUE)))
write.table(cpaai, "cpAAI_matrix.txt", sep='\t', col.names=NA, row.names=T)
# the tree obtained based on the concatenate can be used to order the matrix
nfrootedtree = "yourNewicktreefile"
rootedconcattree = read.tree(nfrootedtree)
rcpaai = cpaai[rootedconcattree$tip.label,rootedconcattree$tip.label]
write.table(rcpaai, "orderred_cpAAI_matrix.txt", sep='\t', col.names=NA, row.names=T)Plotting can be done thanks to the phylo.heatmap function from the R package phytools:
library(phytools)
library(RColorBrewer)
pdf(paste(nfrootedtree, "heatmap_cgANI.pdf", sep='.'), width=20, height=14)
phylo.heatmap(rootedconcattree, rcpaai, fsize=.5, colors=c(rev(brewer.pal(8, 'YlGnBu')), brewer.pal(9, 'YlOrRd')))
dev.off()