Sniffles is a SV caller for long reads. It is mainly designed for PacBio reads, but also works on Oxford Nanopore reads. SV are larger events on the genome (e.g. deletions, duplications, insertions, inversions and translocations). Sniffles can detect all of these type and more such as nested SVs (e.g. inversion flanked by deletions or an inverted duplication). Furthermore, Sniffles incorporates multiple auto tuning functions to determine data set depending parameter to reduce the overall risk of falsely infer SVs.

Newest release

Code (.zip)

Code (.tar.gz)

wget https://github.com/fritzsedlazeck/Sniffles/archive/master.tar.gz -O Sniffles.tar.gz
tar xzvf Sniffles.tar.gz
cd Sniffles-master/
mkdir -p build/
cd build/
cmake ..
make

cd ../bin/sniffles*
./sniffles

Note Mac users often have to provide parameters to the cmake command:

cmake -D CMAKE_C_COMPILER=/opt/local/bin/gcc-mp-4.7 -D CMAKE_CXX_COMPILER=/opt/local/bin/g++-mp-4.7 .. 

https://bioconda.github.io/recipes/sniffles/README.html

Make sure you have a sorted bam file either from ngmlr or from bwa. For the later make sure you have used -M parameter for mapping to mark which alignments are primary and which are secondary!

 ./sniffles -m mapped.sort.bam -v output.vcf

This produces an output.vcf file, which includes all SVs that were called by the default parameter. Note you have to adjust the parameters for low coverage cases.

If you are interested in calling SVs across multiple samples: Population calling

The folder test_data includes an example of a small deletion that you can run to check the correct installation of Sniffles on your system. See the README in that folder for instructions and approximately runtime.