SINA - reference based multiple sequence alignment
==================================================

|latest| |Bioconda| |downloads| |TravisCI| |CircleCI| |Read the Docs| |Codecov|

.. |latest| image:: https://img.shields.io/github/release/epruesse/SINA/all.svg?label=latest
.. |release| image:: https://img.shields.io/github/release/epruesse/SINA.svg
.. |Bioconda| image:: https://img.shields.io/conda/vn/Bioconda/sina.svg
   :target: https://bioconda.github.io/recipes/sina/README.html
.. |TravisCI| image:: https://img.shields.io/travis/epruesse/SINA.svg?label=build%20(TravisCI)
   :target: https://travis-ci.org/epruesse/SINA
.. |CircleCI| image:: https://img.shields.io/circleci/project/github/epruesse/SINA.svg?label=build%20(CircleCI)
   :target: https://circleci.com/gh/epruesse/SINA
.. |Codecov| image:: https://img.shields.io/codecov/c/github/epruesse/sina.svg
   :target: https://codecov.io/gh/epruesse/SINA
.. |Read the Docs| image:: https://img.shields.io/readthedocs/sina/latest.svg
   :target: https://readthedocs.org/projects/sina/builds
.. |downloads| image:: https://img.shields.io/conda/dn/bioconda/sina.svg?style=flat


SINA aligns nucleotide sequences to match a pre-existing MSA using
a graph based alignment algorithm similar to PoA. The graph approach
allows SINA to incorporate information from many reference sequences
building without blurring highly variable regions. While
pure NAST implementations depend highly on finding a good match in
the reference database, SINA is able to align sequences relatively
distant to references with good quality and will yield a robust result
for query sequences with many close reference.

Features
--------

- Speed. Aligning 100,000 full length rRNA against the SILVA NR takes 40 minutes on a mid-sized 2018 desktop computer. Aligning 1,000,000 V4 amplicons takes about 60 minutes.
- Accuracy. SINA is used to build the SILVA_ SSU and LSU rRNA databases.
- Classification. SINA includes an LCA based classification module.
- ARB. SINA is able to directly read and write ARB_ format files such as distributed by the SILVA_ project.

.. _SILVA: https://www.arb-silva.de
.. _ARB: https://www.arb-home.de

Online Version
--------------

An online version for submitting small batches of sequences is made
available by the SILVA_ project as part of their
`ACT: Alignment, Classification and Tree Service <https://www.arb-silva.de/aligner>`_.
In addition to SINA's alignment and classification stages, ACT allows directly building
phylogenetic trees with RAxML or FastTree from your sequences and (optionally)
additional sequences chosen using SINA's add-neighbors feature.

Installing SINA
---------------

The preferred way to install SINA locally is via `Bioconda <https://bioconda.github.io>`_.
If you have a working Bioconda installation, just run::

    conda create -n sina sina
    conda activate sina

Alternatively, self-contained images are available at
https://github.com/epruesse/SINA/releases. Choose the most recent ``tar.gz``
appropriate for your operating system and unpack::

   tar xf sina-1.7.2-dev-linux.tar.gz
   cd sina-1.7.2-dev
   ./sina


Documentation
-------------

The full documentation is available at https://sina.readthedocs.io.

The algorithm is explained in the paper:

  Elmar Pruesse, Jörg Peplies, Frank Oliver Glöckner; *SINA: Accurate
  high-throughput multiple sequence alignment of ribosomal RNA
  genes.* Bioinformatics 2012; 28 (14): 1823-1829.
  `doi:10.1093/bioinformatics/bts252 <https://doi.org/10.1093/bioinformatics/bts252>`_