text stringlengths 1 330k |
|---|
Format: TXT Size: 58KB Download fileOpen Data |
BioXSD 1.1: Enhanced and optimized XML format |
BioXSD has been developed as a universal XML format for the basic types of bioinformatics data that is in particular suitable to be used with Web services [16]. It models common types of data for which a specialized XML Schema (XSD) has not been widely adopted: biomolecular sequences, alignments, sequence feature records, and references to ontologies and data resources. The BioXSD schema defines formats of data but not formats of particular XML documents, by defining XSD types but no global XML elements. BioXSD types can thus be used according to applications' needs in applications' own XSDs such as those in WSDL files of Web services. |
BioXSD 1.0 type AnnotatedSequence can represent annotations of a biomolecular sequence or genome with any types of positioned or non-positioned features, which can be combined in one record. Although the textual serialization of XML is in general more verbose than a tabular format, already the BioXSD 1.0 has included a number of optimizations compared to traditional feature formats like GFF or BED, thanks to the tree-like structure of XML. These have been mainly: |
• not repeating the reference to a sequence in every feature occurrence |
• not repeating the type of feature in every feature occurrence |
• representing multi-segment and multi-point feature occurrences in one feature-occurrence element |
The goal of BioXSD version 1.1 has been to further improve the expressiveness of the BioXSD formats and at the same time focus on optimizations of the data size. The successor of BioXSD 1.0 AnnotatedSequence is BioXSD 1.1 type FeatureRecord. BioXSD 1.1 in general allows more types of sequence positions, distinguishing them in the same way as the tabular GTrack format. Sparse positions are segments, points (actual points or insertions), and outer positions. Dense positions have been added: dense points (function) marked-up by < nextPoint/> empty elements; and dense partition or step function marked by < nextPartition max="..."/> elements including the border position where each interval ends. However in contrast to GTrack, the different types of positions can still be freely combined within a FeatureRecord. The representation of all types of sequence positions have been refactored, simplified, and optimized. Another crucial set of optimizations allows specification of the ontologies, databases, and computational tools of interest in a condensed way for a list of feature annotations, so that they do not have to be repeated. Detailed contents of the BioXSD feature record are listed in Table 5. Examples of data represented in BioXSD 1.1 format are available at [21-25]. |
Table 5. The allowed content of a BioXSD FeatureRecord |
There is one slight difference in how the GTrack and BioXSD deal with focus of feature records. GTrack defines one operational focus of a concrete dataset. That is the reason why it allows to specify only one type of track locations and only one value column and one edges column at a time, although other values and edges may still be "hidden" in out-of-focus columns. BioXSD on the other hand allows combining features, types of track positions, values, and interconnections freely without any operational focus. Thus, if a tool consuming BioXSD feature data demands it, a particular operational focus of the data must be supplied by the user. |
Compared to other generic sequence-feature formats, BioXSD allows defining complex, structured meanings of annotations, as well as complex feature data and metadata, or relations. This would not be conveniently possible in a tabular format and takes advantage of the XML. BioXSD types can freely be combined and included within documents, files, or applications' inputs and outputs. They can easily be combined with other XML formats defined in other XSDs, can be extended just like classes in an object-oriented programming language, or further restricted using built-in XSD mechanisms. BioXSD can be validated and parsed by ordinary XML/XSD-handling frameworks. |
It has, however, been problematic to use XML formats for highly voluminous data such as whole-genome annotations. The textual serialization of XML is more verbose compared to a textual tabular format, and even more compared to a bespoke binary format. Many basic XML-handling tools have high runtime demands for computer memory, making parsing of huge XML documents impossible. All these problems are hopefully going to be solved thanks to the recent and long-expected Efficient XML Interchange (EXI) standard by the World Wide Web Consortium [17], together with its growing family of supporting libraries, and tools for streamed XSLT transformations and random-access XPath and XQuery queries. EXI defines the way any XML data or document should be serialized in a standard binary format that will be many times smaller and at the same time faster to access than the textual XML. There is no need to develop one's own bespoke binary encodings and parsers when using EXI, and the data can be programmatically handled transparently, with the same look and feel as the ordinary XML. |
Availability of specifications and supporting tools |
The BioXSD 1.1 XML Schema is available at [26]. BioXSD data can be validated by all the main XML validation tools, and consumed and produced programmatically by the bulk of the common XML/XSD-handling libraries. Further information and documentation are available at [27]. |
A complete specification of the GTrack format version 1.0 is attached as Additional file 1 and is also available from the GTrack website [20]. The website also contains supporting tools for the GTrack format, connected to the Genomic HyperBrowser [10,28]. Table 6 contains an overview of all GTrack-related tools available as webtools. |
Table 6. Overview of the webtools available from the GTrack website [20] |
The GTrack format is maintained by Sveinung Gundersen and the BioXSD format is maintained by Matúš Kalaš. Both formats are licensed under the Creative Commons Attribution-NoDerivs 3.0 Unported License [29]. |
The Genomic HyperBrowser [10,28] is built on top of the Galaxy framework [30,31] and provides a large set of statistical investigations tailored for the specific track types of supplied tracks. In order for such analyses to be efficient, the system uses a binary storage scheme internally. In this scheme, the core informational columns are stored as C vectors directly written to disk. The vector files are then accessed using the NumPy package [32] for Python [33], allowing very efficient vector computations. A linear index of the files is built in order to allow random access to the data. This binary representation is stored in parallel to the files in their original format, and updated automatically as the original files are updated. The implementation is open source and available as part of the HyperBrowser code base under the GPL license, version 3 [34]. As an alternative, the recently published Tabix tool [35] provides fast access to tabular data in compressed form, and works with GTrack files of types Points and Segments, and their derivatives. |
By systematic analysis of informational properties of genomic tracks, we delineated fifteen distinct types of tracks. These track types shed light on the variability of track representations, suggesting that the differences between formats is not only due to preferences and conventions, but also to fundamental differences in the information inherent in different tracks. Furthermore, discerning the informational properties of a track allows the nature of the track to be precisely conveyed, as well as clarifying what represents meaningful analyses on a given track. |
The identification of core informational properties of tracks, as well as a broad survey of various practicalities concerning existing formats, created a basis for the specification of a new format for genomic data: the GTrack format. By allowing precise interpretation, simple parsing, as well as relatively straightforward conversion to several existing formats, we believe that the introduction of this "yet another format" will actually help streamline data representation in the field. Finally, by coordinating the GTrack format with an enhanced and optimized version 1.1 of the BioXSD format, this also aids in unifying tabular and XML-based track representation, while keeping the specific advantages of the two. |
BAM: Binary Alignment/Map format; BED: Browser Extensible Data format; ChIP-seq: Chromatin Immunoprecipitation sequencing; EXI: Efficient XML Interchange; F: function; GFF: General Feature Format; GTF: Gene Transfer Format; GVF: Genome Variation Format; GP: genome partition; P: points; LBP: linked base pairs; LF: linked function; LGP: linked genome partition; LP: linked points; LS: linked segments; LSF: linked step function; LVP: linked valued points; LVS: linked valued segments; S: segments; SAM, Sequence Alignment/Map format; SF: step function; SNP: single nucleotide polymorphisms; URI: Uniform resource identifier; URL: Uniform resource locator; VP: valued points; VS: valued segments; WIG: Wiggle format; WSDL: Web Service Definition Language; WYSIWYG: what you see is what you get; XML: Extensible Markup Language; XSD: XML Schema Definition. |
Authors' contributions |
SG, AF, EH and GKS conceived and developed the ideas on track type distinctions. SG, MK, OA and GKS developed the GTrack specification. SG and GKS wrote the main parts of the paper. MK wrote the parts on XML-based track representation and developed BioXSD 1.1. SG and GKS were involved with the development of GTrack-related tools. All authors read and approved the final manuscript. |
Acknowledgements and funding |
Funding was kindly provided by EMBIO, FUGE, UiO, Helse Sør-Øst, and eSysbio (funded by the Research Council of Norway). This work was performed in association with 'Statistics for Innovation', a Centre for Research-Based Innovation funded by the Research Council of Norway. We thank Kai Trengereid for crucial work in developing the GTrack-related tools, and Inge Jonassen for valuable input on the BioXSD format. We would also like to acknowledge the excellent review work provided by the peer reviewers. These reviews have contributed significantly to the content of this paper. |
Science 2009, 326(5950):289-293. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL |
2. Generic Feature Format version 3 [] webcite |
4. UCSC genome browser data formats [] webcite |
5. Definition of Gene Transfer Format [] webcite |
6. Reese MG, Moore B, Batchelor C, Salas F, Cunningham F, Marth GT, Stein L, Flicek P, Yandell M, Eilbeck K: A standard variation file format for human genome sequences. |
Genome Biol 2010, 11(8):R88. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL |
7. Liu F, Tostesen E, Sundet JK, Jenssen TK, Bock C, Jerstad GI, Thilly WG, Hovig E: The human genomic melting map. |
PLoS Comput Biol 2007., 3(5) OpenURL |
8. Definition of Wiggle Track Format [] webcite |
9. The Sequence Ontology [] webcite |
Genome Biol 2010, 11(12):R121. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL |
13. Web services provided by the Center for Biological Sequence analysis (CBS), Technical University of Denmark [] webcite |
14. UniProt C: The Universal Protein Resource (UniProt) in 2010. |
Nucleic Acids Res 2010, (38 Database Issue):D142-8. OpenURL |
15. Gould CM, Diella F, Via A, Puntervoll P, Gemund C, Chabanis-Davidson S, Michael S, Sayadi A, Bryne JC, Chica C, Seiler M, Davey NE, Haslam N, Weatheritt RJ, Budd A, Hughes T, Pas J, Rychlewski L, Trave G, Aasland R, Helmer-Citterich M, Linding R, Gibson TJ: ELM: the status of the 2010 eukaryotic linear motif resource. |
Nucleic Acids Res 2010, (38 Database Issue):D167-80. OpenURL |
16. Kalas M, Puntervoll P, Joseph A, Bartaseviciute E, Topfer A, Venkataraman P, Pettifer S, Bryne JC, Ison J, Blanchet C, Rapacki K, Jonassen I: BioXSD: the common data-exchange format for everyday bioinformatics web services. |
Bioinformatics 2010, 26(18):i540-6. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL |
17. Efficient XML Interchange (EXI) Format 1.0 [] webcite |
Bioinformatics 2010, 26(17):2204-2207. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL |
BMC Bioinformatics 2008, 9:523. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL |
20. GTrack [] webcite |
21. BioXSD example 1 [] webcite |
22. BioXSD example 2 [] webcite |
23. BioXSD example 3 [] webcite |
24. BioXSD example 4 [] webcite |
25. BioXSD example 5 [] webcite |
26. Definition of BioXSD version 1.1 [] webcite |
27. [] webcite |
28. The Genomic HyperBrowser [] webcite |
29. Creative Commons Attribution-NoDerivs 3.0 Unported License (CC BY-ND 3.0) [] webcite |
Genome Biol 2010, 11(8):R86. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL |
Curr Protoc Mol Biol 2010, 19:-21. |
Unit 19.10.1 |
32. Oliphant T: Guide to NumPy. Trelgol Trelgol Publishing; 2006. OpenURL |
33. The Python Language Reference [] webcite |
34. GNU General Public License, version 3 [] webcite |
35. Li H: Tabix: fast retrieval of sequence features from generic TAB-delimited files. |
Bioinformatics 2011, 27(5):718-719. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL |
36. Affymetrix CNT File Format [http:/ / SNP_Variation/ Manual/ svs7/ affymetrix_cnt_file_format.html] webcite |
37. VCF (Variant Call Format) version 4.1 [http:/ / wiki/ Analysis/ Variant%20Call%20Format/ vcf-variant-call-format-version-41] webcite |
38. The SAM Format Specification (v1.4-r985) [] webcite |
39. BioHDF [] webcite |
40. FASTA [] webcite |
41. Hoffman MM, Buske OJ, Noble WS: The Genomedata format for storing large-scale functional genomics data. |
Bioinformatics 2010, 26(11):1458-1459. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL |
Re: UX Redesign: Dual Boots / Resize issues / Saving KS |
2011/6/22 Máirín Duffy <duffy fedoraproject org> |
There isn't much we can do for them; they really need to run chkdsk on |
their own. One resource we do have at our disposal is ntfsfix, which is |
a utility that can fix common errors, |
It fixes common errors and then forces windows to run a full check in the next boot. |
but the downside to this is that |
if users press any key while it runs, it'll skip the filesystem check |
operation, rendering the whole process useless. |
That's not a bug we can fix, that's just what windows does... |
So we'll need to lean on |
them and make sure we document well the chkdsk process. Will found this |
following documentation on shrinking partitions in Windows: |
Which is useless, because: |
Windows partitioning tools can't resize the main OS drive (drive C). |
The UI is ugly and not usable, users will have problems with it. |
Resize issues |
screen map. |
There's a complication that resizing introduces. The designers would |
case is not reversible at this time. |
dry-run version. |
A dry-run version of resize would need two components: |
(1) Dry run of the resize to determine how much space we'd get |
(2) A dry run of the RPM transaction to determine how much space we'll |
Well, we can't do rpm transaction dry run in anaconda for network installs before we download everything. |
My idea was, simply adding a new field to comps with auto-generated size requirement for the specific group defaults that will be generated on compose time. (I'm not sure Infra will like this idea though). |
If we want to allow individual package selection, the filed should have the installed size for each package in comps (and be in the package definition, not group definition), not for the whole group. |
There's a complication with upgrades here too, especially during |
As I've mentioned, upgrades doesn't need re-sizing, (and iirc you *can't* change the partition scheme during upgrades). |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.