RESUMO
This paper describes the application of a probabilistic neural network (PNN) to the classification of normal human chromosomes. The inputs to the network are 30 different features extracted from each chromosome in digitized images of metaphase spreads. The output is 1 of 24 different classes of chromosomes (the 22 autosomes plus the sex chromosomes X and Y). An updating procedure was implemented to take advantage of the fact that in a normal somatic cell only two chromosomes can be assigned to each class. The network has been tested using the Copenhagen, Edinburgh, and Philadelphia databases of digitized images of human chromosomes. The recognition rates achieved in this study are superior to those reported using either the maximum likelihood or back propagation neural network techniques.
Assuntos
Cromossomos Humanos/classificação , Redes Neurais de Computação , Bases de Dados Factuais , Humanos , Cariotipagem , ProbabilidadeRESUMO
We survey and discuss issues required of intelligent systems to support research efforts in locus mapping. In particular we focus on the issues of order, on how one can automate the reasoning processes of ordering, and the database structures required to support orders, including ambiguity and uncertainty. We conclude with a summary of work to be done.
Assuntos
Inteligência Artificial , Mapeamento Cromossômico/métodos , Bases de Dados Factuais , Animais , Simulação por Computador , Humanos , Camundongos , Modelos GenéticosRESUMO
Bibliographic references are an important part of databases and information resources used by clinicians and biomedical researchers. In addition to the obvious clerical advantages of standard references, bibliographic references can also be used as links to related items in different data sets. This paper describes an effort that involved matching references from a variety of disparate databases to canonical MEDLINE references. The references matched were those involved in a database unification project which is part of the Mouse Genome Informatics effort at The Jackson Laboratory. Software was developed to take advantage of a commercially available retrieval engine which accesses MEDLINE on CD-ROM disks. The software permits client programs on UNIX/C, and potentially other environments, to access unabridged MEDLINE via networks supporting the TCP/IP protocols. The matching process described can be used as a model for similar efforts with different research or clinical data sets, as well as different hardware or software environments.