Today, many genetics laboratories have access to modern capillary DNA sequencing machines, such as the ABI PRISM 3100, 3700 or 3730. These machines generate vast amounts of raw sequence data with little user intervention. Consequently, the amount of data to be analyzed has expanded and the bottleneck now is the analysis capacity. Data analysis capacity can be increased by higher levels of automation. Investments in infrastructures to process the raw sequencing data in sophisticated but rigid pipelines might be justified for larger laboratories and larger projects but might be too costly for smaller laboratories. In addition, rigid pipelines are too impractical if different projects share run-time on the same machine while requiring (slightly) different analysis procedures (e.g. vector trimming is needed in plasmid sequencing, but needless when sequencing PCR products).

Nucleotide sequence analysis can be performed with a variety of software tools. Although the number of console and web-based software tools has grown rapidly, the routine use of data input, output and storage may be inconvenient. Furthermore, for performing a series of analyses with different software tools, the sequence data need to be reformatted to the required data structure. Alternatively, sophisticated software suites that provide an integrated environment often are expensive.

Several of the available software solutions are designed to facilitate automated DNA sequence analysis at low cost. Well-known solutions are the Staden package and Bioperl.

The Staden Package contains pregap4 and gap4, full-featured applications with an intuitive graphical user interface [1]. These programs handle a list of raw sequence reads method-by-method. The programs in the Staden Package typically require a degree of user intervention and thus hands-on time.

Alternatively, Bioperl is a group of perl modules describing many genetics and genomics concepts [2]. For example, it includes the Bio::Seq::SeqWithQuality object that provides some of the basic properties of a raw sequence (i.e. its nucleotide sequence and quality values); the Bio::Tools::Primer3 object provides methods to work with primer3 input and output. However, to build custom DNA sequencing data pipelines, basic programming skills are needed to combine all these modules.

Smaller laboratory sites, however, often need to implement versatile pipelines that can be adjusted for any research question that suits the project best; at the same time, they often also do not have dedicated programmers available.

Although (semi-)automated procedures have been published by other groups [3, 4], these are mostly focused on one particular pipeline and environment.

Here, we present POSA, a set of two new perl objects (Read.pm and Contig.pm) that describe a raw sequence and a Phrap contig in detail and are easily implemented in perl-based pipelines. Because these objects provide building blocks for sequencing data analysis pipelines and the actual pipelines are built using perl-scripts, the POSA objects can be used in very diverse settings.

The POSA source code is entirely coded in object-oriented Perl and consists of two objects: Read.pm and Contig.pm. In general, there are two important concepts associated with objects: methods (built-in procedures that can be performed on the object) and properties (describing some of the characteristics of the object). Most methods in the objects rely on the availability of other third-party programs (see Dependencies). Basically, POSA provides a wrapper around these programs and provides easy design and implementation of these programs in automated data analysis. The Read.pm object describes a DNA sequence trace and includes methods for data import from a variety of formats. It relies on Phred [5, 6] for import and interpretation of raw sequence data. The original trace data are stored in binary (scf) format within the object. Other methods of Read.pm use modules of the Staden Package [1], such as qclip and vector_clip (if installed). Properties of Read.pm include e.g. the DNA sequence, quality scores, template and vector names and read direction.

The Contig.pm object contains a method to assemble contigs of reads using the Phrap program [6]. The object typically is created based on a list of Read.pm objects and can be exported as alignments or screened for polymorphisms using PolyPhred [7].

Both the Read.pm and Contig.pm objects were designed with flexibility in mind. To allow a (virtually) unlimited amount of data to be processed, the perl scripts using these objects work sequence-by-sequence rather than method-by-method. Typically, these objects are called from straightforward perl scripts that outline the analysis steps to be performed. Example scripts using the objects can be accessed from the download website. An example of a script and output using the two objects to process a set of reads and annotate sequence polymorphisms from the assembled contig is given in Figure 1 and Figure 2.

POSA was developed with perl 5.6.1 and tested on a SuSE linux 8.1 system for abi-files from the ABI PRISM 377 DNA Sequencer and 3100 Genetic Analyzer (Applied Biosystems). Phred, Phrap and PolyPhred versions were 0.000925.c, 0.990329 and 4.05, respectively.

POSA describes a DNA sequence read and a Phrap contig assembly in detail. These objects allow a flexible and easy setup of perl-based pipelines to handle DNA sequencing data, including generating primer3 input and automated SNP discovery, while requiring only little programming skills.

Project name: POSA

Project home page: http://posa.sourceforge.net

Operating system: platform independent Programming language: Perl 5.6.1

License: Artistic License (Open Source)