On Monday 08 January 2007 14:55, Andreas Tille wrote: > On Mon, 8 Jan 2007, Steffen Moeller wrote: [..] > >> field::medicine:practice > > > > fine, I would not mind "information systems" instead of practice, though. > > Or we should distinguish between practicioners, dentists and hospital > systems. It sounds at least reasonable to distiguish between small > practices and hospitals because the software in these fields will be > most probably not be identical. I don't know, so I must admit. > > > fields::biology::sequence > > fields::biology::sequence::dna > > fields::biology::sequence::rna > > fields::biology::sequence::protein all the above are covered by some alignment pretty printer. Gene prediction programs only work on DNA, I think I have seen the dynamite package in this context beyond its ITP stage. If one considers to tag ITPs, too, then there is little doubt that the distinction is required. Protein threaders are not _yet_ in Debian, but they work only on protein sequences, perform alignments and use protein structures. Polyxmass could go here, too, I think. > > fields::biology::structure Rasmol > > fields::biology::interaction The pkg-bioc KEGG library > > fields::biology::genomics The pkg-bioc affy libraries and most other bits, like GO. > > fields::biology::proteomics polyxmass and some libraries in pkg-bioc > > fields::biology::metabolomics well ... don't know at the moment. It should be there, though, maybe also polyxmass or the KEGG library from above. The fields::biology::system is missing, indeed. I have a package for the SBML library locally (and an ITP for it, or had it) and also once packaged E-cell. > Could you please add at least one package to each of these? I did above. While desireable, I do not think that it is ultimately required to have a package available for everything one has a tag for. > > Ah, right and on formats. The following come to mind: > > works-with-format::biology::fasta > > works-with-format::biology::aln > > works-with-format::biology::pdb > > works-with-format::biology::embl (EMBL-format works both for protein and > > nucleotide sequences) > Ahhh, this is a VERY important hint. Good. But how about becoming more specific, like: works-with-format::biology::affy::dat works-with-format::biology::affy::cel Very useful for very few people, but for the masses? Getting back to Charles' question about other ontologies, there are plenty: http://www.bioontology.org/ncbo/faces/pages/ontology_list.xhtml (did I mention I once did a package for the OBO ontology editor :-/ ?) but not really one on bioinformatics tools. The MyGrid consortium came up with an ontology of services, which seems within reach of ours, if we only distinguish between having the tags and having the programs for it http://www.mygrid.org.uk/ontology/myGridDomainOntology.rdfs Although less directly fitting, the GO Evidence Codes may stimulate discussions. The developers have thought of a bioinformatics ontology themselves in order to better characterise results derived from in silico experiments. Useful as a reference are certainly the Medical Subject Headings (MeSH) on http://www.nlm.nih.gov/mesh/meshhome.html and the tree on http://www.nlm.nih.gov/mesh/2007/MeSHtree.L.html. Many greetings Steffen
Attachment:
pgpifuCbjtCSO.pgp
Description: PGP signature