BUSCO, like Cegma, is a special tool for the field of "completeness assessment". This concerns genome assemblies, particularly ones generated de-novo, when by concentrating on a core set of genes, one can estimate how complete the assembly is by the number of the these core genes that the assembly has managed to recover.
BUSCO actually stands for "Benchmarking Universal Single-Copy Orthologs" and labels itself as a quality measure of the assembly. Busco also means "I search" in the Spanish, Galician and Portuguese languages, in which the authors find satisfaction, as the broad goal of the tool is one of a quest for quality.
BUSCO can work closely with augustus, even as far as undertaking retraining (for a species). However, take note:
"Write access to the Augustus installation directory is necessary for retraining the gene finder", so retraining is probably best carried out by the sysadmin under "root" user.
BUSCO is primarily a python application and though it can - apparently - work with version 2 of python, it is version 3 that is recommended and implemented on marvin.
The broad BUSCO process is as follows:
- identification of candidate regions from the genome to be assessed with tBLASTn searches using BUSCO consensus sequences.
- Gene structure prediction using Augustus with BUSCO block profiles.
- These predicted genes, or all genes from an annotated gene set or transcriptome, are assessed using HMMER and lineage-
specific BUSCO profiles to classify matches as "complete", "duplicated", or "fragmented", or when there are no matches, as "missing".
After the first blastn, augustus is invoked as follows:
augustus --proteinprofile=example/prfl/BUSCO_7.prfl --predictionStart=163394 --predictionEnd=174110 --species=fly "sampleasroo2_.temp" > ./run_asroo2//augustus/BUSCO_7.out.1 2>/dev/null
loading the module
module load BUSCO
is enough, as all BUSCO's dependencies (python/3.4, augustus/3.2.2, hmmer/3.1b2, EMBOSS/6.6.0) will also be loaded at the same time.
The main BUSCO executable is a python script called. Becuase of the loaded modules this should run under python3.4 which is the recommended version of python for Busco.
BUSCO has the following three modes
- Genome assembly assessment, i.e. -m genome
- Transcriptome assembly assessment i.e. -m trans
- Gene set assessment i.e. -m ogs
typical usage command lines
After loading the module, the template is as follows:
BUSCO.py -o NAME -in ASSEMBLY -l LINEAGE -m genome
- -o is simply for giving a name to an output directory where various output files will be stored. run_ will be put in front of this name.
- -in is you data input, which for BUSCO is the assembled genome.
- -l is the lineage, which can be one of the following: $METAZOA_LIN, $EUKARYOTA_LIN, $BACTERIA_LIN, $ARTHROPODA_LIN, $VERTEBRATA_LIN, $FUNGI_LIN: these are variables telling BUSCO where to look.
- -m as already explained this is the mode.
So, a typical command on a bacteria assembly would run like this:
BUSCO.py -in scaffolds.fasta -o busc0 -l $BACTERIA_LIN -m genome
typical usage queue job script
Here is how we might write a job script to run a BUSCO job on a bacteria lineage, with 4 running processes:
#!/bin/bash #$ -cwd #$ -j y #$ -S /bin/bash #$ -V #$ -q highmemory.q #$ -pe multi 4 EXPECTED_ARGS=2 # we should feed two arguments to this script # some quick "argument accounting" if [ $# -ne $EXPECTED_ARGS ]; then echo "Sorry, not run because you must supply this script with $EXPECTED_ARGS arguments" echo "They are 1) the input genome assembly 2) name for your output directory (will be prefixed with \"run_\")" exit fi module load BUSCO BUSCO.py -in $1 -o $2 -c $NSLOTS -l $BACTERIA_LIN -m genome
NOTE the usage of the -c (CPUs) option here, it specified how many CPUs/process/threads BUSCO will run with, alllowing parallelisation and faster processing. Its value is $NSLOTS with rfere to the -pe multi 4 line appearing higher up. i.e. BUSCO will run with 4 parallel threads in this case.
If we named this script, say, runbuscbact.sh, then we could run it like so (not forgetting to give the input assembly file, with the path to the directory in which it may be found, and our chosen output folder name which will be prefixed with run_):
qsub runbuscbact.sh 504302_trimmoed2/scaffolds.fasta out0
If the gb subdirectory in output directory is empty that means that the augustus section of the BUSCO procedure did not run properly.
writeable augustus directory
There is a certain error, which, because it only refers to the re-training operation, can be safely ignored (unless when wanting to undertake re-training of course)
Error: Cannot write to Augustus directory, please make sure you have write permissions to /usr/local/Modules/modulefiles/tools/augustus/3.2.2/config
This was corrected by creating a world writeable directory in shelf (scratch) filesystem.