Simulate Structural Variations and SNPs for artificial DNA templates. The DNA templates are generated simply with a pseudo-random number generator, and is not guided by biological significance. The tool is provided with the intention of creating a validation set of variants to test current structural variation detection tools.
The recommended method is to install via pip or git clone. You can also get the
source distribution from pypi. pip will install all the necessary dependencies for you.
$ pip install simulatesv
or
$ git clone [email protected]:lioumens/simulatesv.git
$ python setup.py install
This script runs with either Python 2 or 3, and requires the Python package numpy to run.
If you install via pip, the script will be available to run as simulatesv.py, or python simulatesv.py if you installed via git clone.
$ python simulatesv.py
This will run the program with all defaults, making one 50 kbp reference genome and three simulated genomes in the current working directory. (total 7 files)
.
├── reference.fna # randomly generated dna template on which mutations are made
├── sim_0.fna # first simulated genome based off reference with SNPs and SVs
├── sim_1.fna # 2nd ...
├── sim_2.fna # 3rd ...
├── changes_0.txt # log of what changes were made to the reference genome to create the first simulated genome
├── changes_1.txt # 2nd change log ...
└── changes_2.txt # 3rd change log ...
You may add a seed number to make the simulation reproducible. We can also specify the number of simulated genomes to make and the length of the generated reference.
$ python simulatesv.py --seed 10 --number 5 --length 60000
If you have your own genome template that you would like to simulate variants
from, you can specify the FASTA file with the -t FILE option. You can also
modify the frequency of each variant as well as their size.
See the Command Reference below for more options.
$ python simulatesv.py -t reference.fna -n 20 --snp-error-rate .005 --largest-insertion-size 200
| type | ref_idx | alt_idx | size | ref | alt |
|---|---|---|---|---|---|
| SNP | 22 | 22 | 1 | A | G |
| INS | 3327 | 3002 | 17 | . | CAACTACTAATCCACCA |
| TRANS | 30292 | 2000 | 12 | AATCGCGCGACT | AATCGCGCGACT |
The changes file logs what changes are made to the reference sequence. A "."
marks that column as a null dna sequence. The file is sorted by ref_idx.
| column | description |
|---|---|
| type | type of structural variation |
| ref_idx | index in reference of ref bases |
| alt_idx | index in simulated genome of alt bases |
| ref | bases in reference genome |
| alt | bases in simulated genome |
-
If you are simulating a small DNA template (< 500), make sure that SV option sizes are less than the total size of your genomes, otherwise you may run into unknown issues.
-
Mutation rates for SVs and SNPs are approximate. A binomial model is used to determine the actual number of mutations based off of the given error rate.
-
There is a subtle difference for the meanings of the
idxcolumns between indels and translocations for the changes file. For translocations, theidxwill track the dna sequence, and will not track the null sequences of where the sequence was deleted or inserted. That is, if you think of a translocation as one deletion and one insertion, there will be two columns tracking theref_idxof the insertion and twoalt_idxfor the deletion.
usage: simulate.py [-h] [-n N] [-s SEED] [-l N | -t FILE] [-b BASE] [-o FILE]
[-c BASE] [-se ERR] [-ie ERR] [-de ERR] [-lis N] [-lds N]
[-sis N] [-sds N] [-te ERR] [-lts N] [-sts N] [-ce ERR]
[-lcn N] [-scn N] [-lcs N] [-scs N]
Simulate DNA template sequences with known SNPs and structural variations.
optional arguments:
-h, --help show this help message and exit
-n N, --number N Number of simulated genome sequences to generate.
(default=3)
-s SEED, --seed SEED Pseudorandom number generator seed. Set number to
reproduce genomes and changes.
-l N, --length N Total length of reference genome sequence.
(default=50000)
-t FILE, --template FILE
Template to use as a base for generating SNP/SV in
Fasta format.
Output File Options:
-b BASE, --basename BASE
Basename of the simulated fake sequences.
(default=sim_)
-o FILE, --output FILE
Reference file for the mutated template
(default=reference.fna)
-c BASE, --changes BASE
Basename of the changes files (default=changes_)
SNP Options:
-se ERR, --snp-error-rate ERR
Error rate for SNPs (default=.001)
Indel Options:
-ie ERR, --insertion-error-rate ERR
Error rate for insertions (default=.0001)
-de ERR, --deletion-error-rate ERR
Error rate for deletions (default=.0001)
-lis N, --largest-insertion-size N
Largest insertion size (default=500)
-lds N, --largest-deletion-size N
Largest deletion size (default=500)
-sis N, --smallest-insertion-size N
Smallest insertion size (default=1)
-sds N, --smallest-deletion-size N
Smallest deletion size(default=1)
Trans Options:
-te ERR, --trans-error-rate ERR
Error rate for translocations (default=.0001)
-lts N, --largest-trans-size N
Largest translocation size (default=500)
-sts N, --smallest-trans-size N
Smallest translocation size (default=10)
CNV Options:
-ce ERR, --cnv-error-rate ERR
Error rate for CNVs (default=.0001)
-lcn N, --largest-cnv-number N
Maximum number of times the CNV can repeat
-scn N, --smallest-cnv-number N
Minimum number of times the CNV will repeat
-lcs N, --largest-cnv-size N
Largest size for CNVs (default=300)
-scs N, --smallest-cnv-size N
Smallest size for CNVs (default=2)
Bug reports, feature requests, and pull requests are encouraged!
MIT