Read Me
LoFreq: A sequence-quality aware, ultra-sensitive variant caller for
high-throughput sequencing data
INSTALLATION
------------
See INSTALL for generic installation instructions.
You will need a C compiler, Python 2.7 including header files and zlib
including header files to install LoFreq.
The following will build and install LoFreq:
$ ./configure
$ make install
You will likely need to use sudo for the last step. If you however
don't have admin rights or want to install LoFreq into a non-standard
directory use the --prefix flag for configure. You will then very
likely have to adjust your PATH and/or PYTHONPATH environment variable
as well. An example:
$ ./configure --prefix $HOME/local/
$ make install
$ export PYTHONPATH=$PYTHONPATH:$HOME/local/lib/python2.7/site-packages/
$ export PATH=$PATH:$HOME/local/bin
USAGE
-----
All script-names start with lofreq_ and print a simple help message
when invoked with -h.
For a full description see https://sourceforge.net/p/lofreq/
and also read the publication (www.ncbi.nlm.nih.gov/pubmed/23066108)
VERSION HISTORY
---------------
0.6.1 (2013-05-30):
- Added option -A|--anomalous-pairs-allowed which will take reads
into account whose mate is unmapped
- Added --force option to lofreq_snpcaller.py and lofreq_filter.py
which allows forced overwriting of output files
0.6.0 (2013-03-26):
- Got rid of Biopython dependency
- Default bonf factor reverted to auto instead of auto-ign-zero-cov
- New option --cons-as-ref for cases where you want to call cosnensus
bases per position and then call SNVs instead of calling SNVs against
the given reference
- Added CONSVAR INFO field to vcf
- Now user supplied CFLAGS and LDFLAGS are passed down to Python
extension building as well
0.5.0 (2013-01-25):
- Base-call (P_bq) and mapping qualities (P_mq) are now joined by default:
P_joined = P_mq + (1-P_mq)*P_bq
- GNU autotools now used for compilation and installation
- Included helper scripts to create a stringent, recalibrated mapping
(bwa_unique.sh and base_qual_calib_wrapper.sh)
- Incorporated a modified version of samtools mpileup
- Fixed mixup between --bonf options auto-ign-zero-cov and auto. The
former is now default (needs extra parsing of BAM
- Added script lofreq_alnoffset.py which makes comparison of SNV
calls made on different coordinate systems / against different
reference sequences easier
- lofreq_varpos_to_vcf.py now easier to use (takes BAM as input)
0.4.0 (2012-11-08):
- Fixed overflow for very large Bonferroni factors (relevant for auto
Bonferroni on WGS of HG for example)
- Reported base-counts and frequencies are now before filtering with
the exception of Q<3 which is always removed
- Added Bonferroni option 'auto-ign-zero-cov'
- Added --force option for forced overwriting of files
0.3.2 (2012-10-17):
- Bonferroni factor now determined automatically by default in
lofreq_snpcaller.py
- Check for samtools during installation
- lofreq_uniq_pipeline.py summarizes result files
0.3.1 (2012-10-10):
- Made LoFreq chromosome aware
- Samtools pileup is now generated internally
- Added LoFreq uniq pipeline
- Added support for region bed files
- Added more helper scripts
0.2 (2012-07-12):
- Used for producing results in the Wilm et al. (2012)
- Optimal implementation of LoFreq-Q as described in publication
- SNV calls always against consensus, which is determined on-the-fly
- If consensus differs from ref, we call it consensus-variant
0.1 (2011-11-10):
- First public release. Initially hosted on github.