seqan / seqan

SeqAn's official repository.

SeqAn - The Library for Sequence Analysis

NOTE SeqAn3 is out and hosted in a different repository
We recommend using SeqAn3 for new applications.

What Is SeqAn?

SeqAn is an open source C++ library of efficient algorithms and data structures for the analysis of sequences with the focus on biological data. Our library applies a unique generic design that guarantees high performance, generality, extensibility, and integration with other libraries. SeqAn is easy to use and simplifies the development of new software tools with a minimal loss of performance.

License

The SeqAn library itself, the tests and demos are licensed under the very permissive 3-clause BSD License. The licenses for the applications themselves can be found in the LICENSE files.

Prerequisites

Older compiler versions might work but are neither supported nor tested.

Linux, macOS, FreeBSD

• GCC ≥ 11
• Clang/LLVM ≥ 15
• Intel oneAPI C++ Compiler 2024.0.2 (IntelLLVM)

Windows

• Visual C++ ≥ 17.0 / Visual Studio ≥ 2022

Architecture support

• Intel/AMD platforms, including optimisations for modern instruction sets (POPCNT, SSE4, AVX2, AVX512)
• All Debian release architectures supported, including most ARM and all PowerPC platforms.

Build system

• To build tests, demos, and official SeqAn applications you also need CMake ≥ 3.12.

Some official applications might have additional requirements or only work on a subset of platforms.

Documentation Resources

• Getting Started
• Manual
• Tutorial
• How-Tos
• API Documentation (stable)

Contact

• Mailing List
• GitHub Project (issues, source code)

seqan's Issues

goDown() a std::string for FMIndex iterator does not compile

Suppose I have an iterator, it, over some index and I wish to descend it using goDown(it, x) where x is some string. x could be a seqan string or infix or a std::string. For an ESA or WOTD index this all work just fine. For a FMIndex only the seqan types work. The std::string version fails to compile. See the code below:

I'm using gcc 4.8.2 with the latest git version of seqan.

#include <seqan/index.h>

using namespace seqan;

typedef String< Dna5 > string_t;
typedef StringSet< string_t > string_set_t;



template< typename Iterator >
void
descend( Iterator i ) {
    std::cout << representative( i ) << "\n";
    if( goDown( i ) ) {
        while( true ) {
            descend( i );
            if( ! goRight( i ) ) {
                break;
            }
        }
    }
}


template< typename Index >
void
test_index( Index index ) {
    typename Iterator< Index, TopDown<> >::Type iterator( index );
    descend( iterator );
    goDown( iterator, string_t( "AC" ) );
    goDown( iterator, 'A' );
    goDown( iterator, "AC" );
    goDown( iterator, std::string( "AC" ) ); //fails to compile for FMIndex
}


int main( int argc, char * argv[] ) {
    string_set_t seqs;
    appendValue( seqs, "AC" );
    appendValue( seqs, "A" );
    appendValue( seqs, "ACGT" );
    test_index( Index< string_set_t, IndexEsa<> >( seqs ) );
    test_index( Index< string_set_t, IndexWotd<> >( seqs ) );
    test_index( Index< string_set_t, FMIndex<> >( seqs ) );
    return 0;
}

Scoring algorithm different from literature

I'd like to revive http://trac.seqan.de/ticket/1091

I have some code in the pipeline that works around this, but I would have to add publicly documented functions that explain our behaviour. And I don't see why we shouldn't just change it.

Cannot use on cygwin

Hi. Thanks for your great work on seqan.

I am trying to use seqan on cygwin, but everythings fails to build because seqan/system.h includes aio.h, which is not available on cygwin.

Is there any way to make it work? Probably some macro to disable asynchronous IO?

Thanks in advance.

arg_parse_doc.h needs to be documented

There is no dox docu here, even though there are many doc docu entries.

Bug in VCF output

For multiple reference sequences, the writeRecord(Vcf) always writes the ID of the first reference, even for variants on other references. So I think there is a bug in the

streamWriteBlock(stream, &(_vcfIOContext.sequenceNames)[record.rID][0],
length((_vcfIOContext.sequenceNames)[record.rID]));

part (write_vcf.h, line 161) because the function writes the same ID for different record.rID values:

correct ID:
rId 1
Id gi|81239530|gb|CP000034.1|

output within writeRecord:
rID 1
ID gi|170079663|ref|NC_010473.1|37
rID 0
ID gi|170079663|ref|NC_010473.1|#""0

cheers, Kathrin

[Docs] Fix multiple documented functions

In class Gaps function iter is listed both under Interface Function Overview and Interface Functions Inherited From. Same issue probably applies to other classes as well.

Documentation of convertMatchesToGlobalAlignment() broken

The documentation of function convertMatchesToGlobalAlignment is broken:

• The function requires a third parameter that is not documented.
• The function only does something if the alignQualityStore is properly sized.

Header guard mismatch

While compiling OpenMS with clang 3.4 a couple of warnings were generated due to mismatches of the header guards in seqan (see below). We actually use SeqAn 1.4.1 but a quick check showed that some of the mismatches are still in master. Please also note that we only use parts of seqan, so that is not a complete list.

<path-to-seqan-installation>/include/seqan/basic/iterator_position.h:37:9: warning: 'SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_x' is used as a header guard here, followed by #define of a different macro [-Wheader-guard]
#ifndef SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_x
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<path-to-seqan-installation>/include/seqan/basic/iterator_position.h:38:9: note: 'SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_' is defined here; did you mean 'SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_x'?
#define SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
        SEQAN_CORE_INCLUDE_SEQAN_BASIC_ITERATOR_POSITION_H_x

<path-to-seqan-installation>/include/seqan/sequence/string_set_dependent_generous.h:38:9: warning: 'SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUS_H_' is used as a header guard here, followed by #define of a different macro
      [-Wheader-guard]
#ifndef SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUS_H_
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<path-to-seqan-installation>/include/seqan/sequence/string_set_dependent_generous.h:39:9: note: 'SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUSH_' is defined here; did you mean 'SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUS_H_'?
#define SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUSH_
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
        SEQAN_SEQUENCE_STRING_SET_DEPENDENT_GENEROUS_H_

<path-to-seqan-installation>/include/seqan/align/alignment_operations.h:37:9: warning: 'SEQAN_CORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_' is used as a header guard here, followed by #define of a different macro
      [-Wheader-guard]
#ifndef SEQAN_CORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<path-to-seqan-installation>/include/seqan/align/alignment_operations.h:38:9: note: 'SEQANCORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_' is defined here; did you mean 'SEQAN_CORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_'?
#define SEQANCORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
        SEQAN_CORE_INCLUDE_SEQAN_ALIGN_ALIGNMENT_OPERATIONS_H_

ValueSize<> for finite alphabet with more than 256 values

On alphabets with more than 256 characters, the ValueSize<> metafunction does not work properly because the return type of this meta function is limited to 8 bit.

#include <seqan/basic.h>
using namespace seqan;
int main()
{  
    // define alphabet with 259 characters
    typedef SimpleType<unsigned, Finite<259> > TAlph;
    std::cout << static_cast<unsigned>(ValueSize<TAlph>::VALUE) << std::endl;  
    return 0;
    // Output: 3
}

Unused typedefs in sequence_journaled_iterator.h

There are warnings about typedefs that are not used. You can see them in the g++-4.8 builds of CDash and (now) in Travis if you pull from develop.

64bit word size breaks the FM index test

It seems the change of word size broke the masai_index test. Please see my proposal in PR #174 (I cannot reopen it).

Backtracking in dev2 but not in dox

Clarify usage of const-ness together with Dependent StringSet.

@weese, we should talk about const-ness together with dependent string sets.

The sets themselves can be modifiable while the strings are const. I think the current code does not allow for all possible/important cases.

Also, getValueById() does not work properly with const string sets of all specializations.

FragmentStore: loading UCSC annotations

Hi! I was trying to write a simple converter from UCSC's files knownGene.txt and knownIsoforms.txt into GTF. But something went wrong and I cannot figure out why.

I've downloaded USCS files from http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/knownGene.txt.gz and http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/knownIsoforms.txt.gz But my code

#include <seqan/store.h>
#include <seqan/sequence.h>
#include <seqan/stream.h>

using namespace seqan;

int main(int argc, char const* argv[])
{
  seqan::Stream<seqan::GZFile> geneStream;
  if (!open(geneStream, "knownGene.txt.gz", "r"))
  {
    std::cerr << "ERROR: Could not open knownGene.txt.gz for reading.\n";
    return 1;
  }

  seqan::Stream<seqan::GZFile> isoformStream;
  if (!open(isoformStream, "knownIsoforms.txt.gz", "r"))
  {
    std::cerr << "ERROR: Could not open knownIsoforms.txt.gz for reading.\n";
    return 1;
  }

  seqan::Stream<seqan::GZFile> gtfStream;
  if (!open(gtfStream, "ucsc.gtf.gz", "w"))
  {
    std::cerr << "ERROR: Could not open ucsc.gtf.gz for writting.\n";
    return 1;
  }

  FragmentStore<> store;

  read(geneStream, store, Ucsc());
  read(isoformStream, store, Ucsc());

  write(gtfStream, store, Gtf());

  return 0;
}

is failing with message:

/data/results/gusev/epigenetics/novo/annotation/build/seqan/include/seqan/sequence/string_base.h:488 Assertion failed : static_cast<TStringPos>(pos) < static_cast<TStringPos>(length(me)) was: 0 >= 0 (Trying to access an element behind the last one!)
Aborted

What can be the problem here? Thank you!

PS: I am using seqan v. 1.4.1.

Documentation of BamHeaderRecord::type is missing.

There appears to be no problem with the dox comment itself but with the documentation system since the @var entry does not appear in the dox sources written to the HTML page.

See:

http://docs.seqan.de/seqan/dev3/class_BamHeaderRecord.html

Parameters of SequenceStream constructor not documented

The Basic IO tutorial mentions that you can specify a file format as a third parameter to the SequenceStream constructor and gives a link to the API doc.
The API mentions the format parameter but there is no info on what it should be for different formats.

Documentation of Alphabets

We found three issues:

1. Return value type of ValueSize<TAlphabet>::VALUE is not documented.
2. T valueSize<T>() should be T1 valueSize<T2>().
3. class AminoAcid: "The amino acids are enumerated from 0 to 15" should be 19, not 15.

Error when trying to do MSA for similar strings

This code I tryed to use in unit testing in my homework cause an error:

   Align<TSequence> align;
   resize(rows(align), 4);
   for(int i = 0; i < 4; ++i)
   {
     assignSource(row(align, i), TSequence("KKKPPPGGF"));
   }
   globalMsaAlignment(align, Blosum62(-1, -1));

Please write at [email protected]

Tutorial corrections

Tutorial First Steps in SeqAn:
The output in the solutions of assignments 5 and 6 and the output in "The Final Result" has too many 0s at the end.

Tutorial Alphabets, Assignment 1, solution:
alphSize should be of type typename ValueSize<TAlphabet>::Type and not typename Size<TAlphabet>::Type.

Tutorial Index Iterators:
"How many wood would a woodchuck chuck." should be "How MUCH wood would a woodchuck chuck.".

Documentation for resizeEdgeMap missing.

Exists in dev2

http://docs.seqan.de/seqan/dev2/files/FUNCTION.resize_Edge_Map.html

Extract CharString from BamTagsDict

The function extractTagValue() always returns false when the tag type is 'Z' (return value of getTagType()). The example given on the documentation page of BamTagsDict outputs

3
AA -> ""
AB -> ""
AC -> 30

and not

3
AA -> "value1"
AB -> "value2"
AC -> 30

Adjust members to naming scheme

There are still some member variables that do not follow the naming scheme:

• BamAlignmentRecord::rNextId -> BamAlignmentRecord::rNextID
• GenomicRegion::seqId -> GenomicRegion::seqID

Also, all fragment store members and related types should be rechecked.

[DOC] File Formats Tags vanished in dox

See http://docs.seqan.de/seqan/dev2/?i=Tag.File%20Format

Windows/Visual Studio Tutorial not building razers2

When walking through the "Getting Started With SeqAn On Windows Using Visual Studio" tutorial, building razers2 does not work. The following error is displayed:

3>C:\Dev\SeqAn\seqan\core\include\seqan/parallel/parallel_sequence.h(77): error C2665: 'seqan::atomicInc' : none of the 2 overloads could convert all the argument types
3> C:\Dev\SeqAn\seqan\core\include\seqan/parallel/parallel_atomic_primitives.h(340): could be 'long seqan::atomicInc(volatile long &)'
3> C:\Dev\SeqAn\seqan\core\include\seqan/parallel/parallel_atomic_primitives.h(341): or 'unsigned long seqan::atomicInc(volatile unsigned long &)'
3> while trying to match the argument list '(volatile unsigned int)'

Also receive the same error for atomicDec.
Using Microosft Visual C++ 2010 Express.

Fix documentation of Swift and Pigeonhole Pattern

The documentation of Swift and Pigeonhole patterns could use some love. Although I had previous knowledge of the code and RazerS 3, I had to look around a bit. The following things need improvements:

• signatures and parameters
• we need working examples
• the parametrization with threshold etc. for Swift and error rate for Pigeonhole should be documented

[Docs] Fix Gaps constructor documentation

Class Gaps has constructor of specialisation Gaps<Tseq, AnchorGaps> showing up in documentation.

BamIndex documentation lacking

Also check BAI specialization.

Move class Fragment to module align

The class Fragment<> is currently part of seqan/graph_types.h. However, it should go to the align module.

Shape#hashInit undocumented

The function is used in the Tutorial here but appears to be missing in dox:

https://trac.seqan.de/wiki/Tutorial/IndexQGram

translateFile2GlobalRefId for rNextId in readRecord

I was reading a bam file with a broken header and stumbled into this:

The function readRecord(record, context, stream, Bam()) in bam_io/read_bam.h corrects the rID but not the rNextId with the field translateFile2GlobalRefId of a BamIOContext. It seems like a bug to me to correct only one of the IDs although I do not fully understand what is going on.

Documentation for toCString() missing in dox

It is there in dev2.

beginPos of unmapped reads in BamAlignmentRecord

According to the sam format specification the begin position of an unmapped read is 0:
"POS is set as 0 for an unmapped read without coordinate."

However, when calling the functions

readRecord(record, context, reader, seqan::Sam())
write2(outStream, record, context, seqan::Bam())

the begin position of an unmapped read is being changed from 0 to -inf in the resulting output file.

NameStoreCache need proper documentation

There is some docu, however, it needs to be extended and an example would be good.

BED are 1-based but should be 0-based.

Currently, BED coordinates are interpreted as in GFF (1-based) but the UCSC website specified 0-based coordinates:

http://genome.ucsc.edu/FAQ/FAQformat.html#format1

Header hard to create for VcfStream

Currently, the only way to do this properly is as follows, with using the underscore member _context. There should be a better way to do this.

for (unsigned i = 0; i < numSeqs(faiIndex); ++i)
{
    seqan::CharString contigStr = "<ID=";
    append(contigStr, sequenceName(faiIndex, i));
    append(contigStr, ",length=");
    std::stringstream ss;
    ss << sequenceLength(faiIndex, i);
    append(contigStr, ss.str());
    append(contigStr, ">");
    appendValue(vcfStream.header.headerRecords, seqan::VcfHeaderRecord("contig", contigStr));
    appendName(*vcfStream._context.sequenceNames,
               sequenceName(faiIndex, i),
               vcfStream._context.sequenceNamesCache);
}

Documentation of FileOpenMode is broken

The example section ends before giving an example.

If you look into file/file_interface.h, you can find much more documentation of the different modes than there appears at http://docs.seqan.de/seqan/dev3/?p=FileOpenMode.

Write BamIndex

There is a read function but no write.

Argument parser enum values should be INPUT_FILE and OUTPUT_FILE

At the moment the enum values are called INPUTFILE and OUTPUTFILE but they should better be called INPUT_FILE and OUTPUT_FILE for consistency with the rest of SeqAn.

Fix whitespace formatting issues in dox signature Display

See for example the following file:

http://docs.seqan.de/seqan/dev3/global_function_globalAlignment.html

The signature is displayed on one line but should be displayed on several lines.

Documentation for readUntilTabOrLineBreak missing in dox

Must be added.

SVN not working in Ubuntu Studio 14.04

When following the start up tutorial get an error whilst using svn as follows;

ubuntu-studio@ubuntu-studio:~/Dev$ svn co https://github.com/seqan/seqan/branches/master seqan-trunk

svn: E235000: In file '/build/buildd/subversion-1.8.8/subversion/libsvn_wc/wc_db.c' line 1671: assertion failed (SVN_IS_VALID_REVNUM(changed_rev))
Aborted (core dumped)

Could you please advise how to fix this.

Mason writes out wrong TLEN in simulated SAM output

Is minus one times the actual TLEN.

Need BamTagsDict variant for const CharString objects.

Currently, BamTagsDict requires a non-const wrapped object.

This is too limiting.

Loading index does not include original sequence names

The original names of sequences are not available when you load an index from disk.

I am using the readAll function to load in sequences from a fasta file. I then generate an index based on the StringSet of the previously read in sequences. The index takes in as input the StringSet of sequences but not the StringSet that contains the sequence identifiers. I then generate the necessary fibres and then write the index to disk. I am then able to load the index back into the program from disk but am unable to access the original names of the sequences that are in the index.

is there any way to include the sequence IDs when writing an index out to disk?

Provide call-consensus function for ProfileChar.

ProfileChar should have a public function that takes the consensus (if any). Currently, this is done by casting to the target type with static_cast<> but that could be considered unintuitive. Either document this feature or provide such a function.

Also, such an explicit function could convert to IUPAC in case of ambiguities.

We might also consider to have a function that does a statistical test.

Add Affix Index Structure

Dear SeqAn team

First, I want to thank you for this excellent library.

As part of an implementation of an approximate string matching algorithm by “Seed-And -Extend”, I am looking for an affix-index structure. Such structures should be great for extending a pattern in both ways.

In a naive solution, I could use 2 FM-Indexes, one of the text and the other on the reversed text . First of all, I would iterate my seed in the “backward” FM-Index and then with an iterator on each index, I would extend in a direction or the other. However, at each change of way, it would be necessary to iterate the entire new pattern in the opposite substring index.

This approach does not seem very efficient compared to what Schnattinger et al proposed in their article “Bidirectional Search in a String with Wavelet Trees”(2010). Indeed, at each iteration in a BWT, they update interval in the opposite BWT with a time complexity of O(log(σ)).

Moreover, the naive solution implies that we have two FibreText loaded in memory, which would be quite space consuming.

I think it would be a great opportunity to be able to use an affix index in SeqAn. I unfortunately do not have sufficient technical kwonledges of SeqAn and template programming to be able to implement new structures. I'm wondering if you're planning to implement (or maybe already implementing!) such a structure in the SeqAn library.

Best regards,

Christophe Vroland

VcfStream::header does not turn up in dox documentation

opening FM-Index from disk needs "write" permission

It is impossible to open the FM-Index from the disk without having the "write" permission on the files. This permission seems to be unnecessary.

Mason simulator crashes for invalid combination of fragment length-read length-sd

Example call for crash:
./mason_simulator -ir Ecoli_O157H7.fa -n 1832820 -o Ecoli_O157H7_reads1.fa -or Ecoli_O157H7_reads2.fa --fragment-mean-size 1000 --fragment-size-std-dev 300 --illumina-read-length 150

Simulating Reads:
gi|15829254|ref|NC_002695.1| (allele 1) /home/trappek/src/seqan/trunk/core/include/seqan/basic/basic_exception.h:345 FAILED! (Uncaught exception of type std::runtime_error: Illumina read is too long, increase fragment length)

stack trace:
0 [0x42a24e] seqan::ClassTest::fail() + 0xe
1 [0x4a4912] ../mason_simulator()
2 [0x7f74a94cc856] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x5e856)
3 [0x7f74a94cb919] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x5d919)
4 [0x7f74a94cc4ca] _gxx_personality_v0 + 0x52a
5 [0x7f74a8d547f3] /lib/x86_64-linux-gnu/libgcc_s.so.1(+0xf7f3)
6 [0x7f74a8d54d27] Unwind_Resume + 0x57
7 [0x46419b] ../mason_simulator()
8 [0x465275] ReadSimulatorThread::run(seqan::String<seqan::SimpleType<unsigned char, seqan::Dna5>, seqan::Alloc >&, PositionMap const&, seqan::String<char, seqan::Alloc > const&, seqan::String<seqan::SimpleType<unsigned char, seqan::Dna5>, seqan::Alloc >&, int, int) + 0x695
9 [0x427c50] ../mason_simulator()
10 [0x4657b8] MasonSimulatorApp::_simulateReadsDoSimulation() + 0x398
11 [0x424711] main + 0x731
12 [0x7f74a8781de5] __libc_start_main + 0xf5
13 [0x424ac1] ../mason_simulator()

Aborted (core dumped)

Package Generation for git tags

The package generator currently doesn' t work with our new Repo on github. Please port it to use bit tags instead of svn tags so new binaries like Fiona can be provided easily.