Database Build

Creating a database

  • We create a database called test.

$ expam create -db test
Successfully created directories /Users/seansolari/Documents/Databases/test/phylogeny!
Fresh database configuration generated!
Logs path /Users/seansolari/Documents/Databases/test/logs created!
Made path /Users/seansolari/Documents/Databases/test/database.
Made path /Users/seansolari/Documents/Databases/test/results.

  • This database will be located in the current directory, in a folder called test.

Add reference sequences

We now add sequences to our new database.

Warning

When using custom reference sequences or assembled genomes, ensure as much as possible that these genomes are uncontaminated - contaminated genomes compromise distance estimation for building the reference tree and subsequently may impact classification performance.

  • We have supplied sequences with the source:

    • Reference genomes: ../expam/test/data/sequence/,

    • Reads (for later): ../expam/test/data/reads/.

  • Add these sequences to our database test.

$ expam add -db test -d ~/Documents/expam/test/data/sequences
Added 6 files from /Users/seansolari/Documents/expam/test/data/sequences/.

Note

expam can handle both compressed and uncompressed sequences files and automatically detects how it should open most regular sequence files.

Specify build parameters

  • These are bacterial genomes, so k=31 should work fine.

  • We’ll use 4 processes to speed up the build.

$ expam set -db test -k 31 -n 4

Build a tree for the reference database

  • Expam needs a tree with the phylogenetic relationship between your sequences.

  • We’ll make a tree out of mash distances.

  • sourmash is a portable Python version of mash (you can install it with pip).

  • We’ll use a sketch size of s=1000 to represent each genome, and then compare them.

$ expam set -db test -s 1000

  • We’ll use RapidNJ to make a tree from the sourmash distances (see here to install).

  • We’ll first ensure that sourmash is installed, before running the tree command to build the tree.

$ python3 -m pip install sourmash
$ expam tree -db test --sourmash
...

  • print and match with the following output:

$ expam print -db test
<<< expam configuration file: test >>>

phylogeny       -->     /Users/seansolari/Documents/Databases/test/phylogeny/tree/test.nwk
k               -->     31
n               -->     4
sketch          -->     1000
pile            -->     None

----------------
group name: default
        k               -->     None
        sketch          -->     None
        sequences       -->     6

Build the database

  • Now that we have a distance-tree for our added reference sequences, we can build the database.

$ expam build -db test
Clearing old log files...
Importing phylogeny...
* Initialising node pool...
* Checking for polytomies...
    Polytomy (degree=3) detected! Resolving...
* Finalising index...
Creating LCA matrix...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000008725.1_ASM872v1_genomic.fna.gz...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000007765.2_ASM776v2_genomic.fna.gz...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000005845.2_ASM584v2_genomic.fna.gz...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000006925.2_ASM692v2_genomic.fna.gz...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000006945.2_ASM694v2_genomic.fna.gz...
Extracting sequences from /Users/ssol0002/Documents/Projects/pam/test/data/sequences/GCF_000006765.1_ASM676v1_genomic.fna.gz...
expam: 42.359643852s

PID - 65856 dying...