SVDF

KO | EN

gitlite — search

#javascript #python #typescript #llm #go #machine-learning #deep-learning #docker #macos #cli #pytorch #security #mcp #ios

(★ 12)

SVDF is a long-read-based structural variants caller uses deep learning network.

SVDF Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# SVDF
SVDF is a long-read-based structural variant caller uses deep learning network.

SVDF is able to detect and genotype DEL/INS/DUP/INV/TRA with fast running speed.

Installation
------------

#Install from github (requires Python 3.6.* or newer): installs all dependencies
    git clone https://github.com/coopsor/SVDF
    cd SVDF

Dependencies
------------
- *tensorflow=2.20* 
- *pandas*
- *numpy* 
- *pysam* 
- *scikit-learn=1.2.2*
- *scipy*

Input
-----

SVDF takes sorted and indexed alignment files in BAM format as inputs. And SVDF has been successfully tested on PacBio CLR, PacBio HiFi (CCS) and Oxford Nanopore data and alignment files produced by the read aligners `minimap2 <https://github.com/lh3/minimap2>`_, `pbmm2 <https://github.com/PacificBiosciences/pbmm2/>`_ , `NGMLR <https://github.com/philres/ngmlr>`.

Output
------

SVDF produces SV calls in the Variant Call Format (VCF).

Usage
----------------------
    python svdf.py call ./HG002_PB_70x_RG_HP10XtrioRTG.bam --working_dir ./ -s 10 -t 16

#'--working_dir' the work path of SVDF to store temporary data and output files, and the output files include called vcf file and the log file.
    #'-s' to filter the low-quality SVs. If you are not sure the parameter, you must specify the sequencing data type(--read_type) and read depth(--depth), SVDF will automatically calculate this value.
    #'-t' to specify the number of threads to use, the default is 16.
    #'--mode' to specify the mode of SVDF, the default is 'general', you can use 'sensitive' mode to call cancer genome SVs or achieve high recalls.
    #'--ref' to specify the reference genome file that the long reads were aligned to. For resolving reference and alternative allele sequences of INDELs.
    #'--skip_genotype' to skip the genotype step, the default is False.
    #'--contigs' to specify the chromosomes list to call SVs, the default is all chromosomes.

Contact
-------

If you experience any problems or have suggestions please create an issue or a pull request.

Citation
--------
Heng Hu, Runtian Gao, Wentao Gao, Bo Gao, Zhongjun Jiang, Murong Zhou, Guohua Wang, Tao Jiang, SVDF: enhancing structural variation detect from long-read sequencing via automatic filtering strategies, Briefings in Bioinformatics, Volume 25, Issue 4, July 2024, bbae336, https://doi.org/10.1093/bib/bbae336

License
-------

The project is licensed under the GNU General Public License.

SVDF

SVDF Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

SVDF

SVDF Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README