Revision [1364]

This is an old revision of BioinformaticsGlossaryO made by LingyunWu on 2005-11-28 17:20:05.

 

Bioinformatics Glossary





O


Object Management Group (OMG)(国际对象管理协作组)
A not-for-profit corporation that was formed to promote component-based software by introducing standardized object software. The OMG establishes industry guidelines and detailed object management specifications in order to provide a common framework for application development. Within OMG is a Life Sciences Research group, a consortium representing pharmaceutical companies, academic institutions, software vendors, and hardware vendors who are working together to improve communication and inter-operability among computational resources in life sciences research. See CORBA.

Object-oriented database(面向对象数据库)
Unlike relational databases (see entry), which use a tabular structure, object-oriented databases attempt to model the structure of a given data set as closely as possible. In doing so, object-oriented databases tend to reduce the appearance of duplicated data and the complexity of query structure often found in relational databases.

Odds score(概率/几率值)
The ratio of the likelihoods of two events or outcomes. In sequence alignments and scoring matrices, the odds score for matching two sequence characters is the ratio of the frequency with which the characters are aligned in related sequences divided by the frequency with which those same two characters align by chance alone, given the frequency of occurrence of each in the sequences. Odds scores for a set of individually aligned positions are obtained by multiplying the odds scores for each position. Odds scores are often converted to logarithms to create log odds scores that can be added to obtain the log odds score of a sequence alignment.

OMIM (一种人类遗传疾病数据库)
Online Mendelian Inheritance in Man. Database of genetic diseases with references to molecular medicine, cell biology, biochemistry and clinical details of the diseases.

Optimal alignment(最佳联配)
The highest-scoring alignment found by an algorithm capable of producing multiple solutions. This is the best possible alignment that can be found, given any parameters supplied by the user to the sequence alignment program.

ORF (开放阅读框)
Open Reading Frame. A series of codons (base triplets) which can be translated into a protein. There are six potential reading frames of an unidentifed sequence; TBLASTN (see BLAST) transalates a nucleotide sequence in all six reading frames, into a protein, then attempts to align the results to sequeneces in a protein database, returning the results as a nucleotide sequence. The most likely reading frame can be identified using on-line software (e.g. ORF Finder).

Orthologous(直系同源)
Homologous sequences in different species that arose from a common ancestral gene during speciation; may or may not be responsible for a similar function. A pair of genes found in two species are orthologous when the encoded proteins are 60-80% identical in an alignment. The proteins almost certainly have the same three-dimensional structure, domain structure, and biological function, and the encoding genes have originated from a common ancestor gene at an earlier evolutionary time. Two orthologs 1 and II in genomes A and B, respectively, may be identified when the complete genomes of two species are available: (1) in a database similarity search of all of the proteome of B using I as a query, II is the best hit found, and (2) I is the best hit when 11 is used as a query of the proteome of B. The best hit is the database sequence with the highest expect value (E). Orthology is also predicted by a very close phylogenetic relationship between sequences or by a cluster analysis. Compare to Paralogs. See also Cluster analysis.

Output layer(输出层)
The final layer of a neural network in which signals from lower levels in the network are input into output states where they are weighted and summed togive an outpu t signal. For example, the output signal might be the prediction of one type of protein secondary structure for the central amino acid in a sequence window.

Overfitting
Can occur when using a learning algorithm to train a model such as a neural net or hid-den Markov model. Overfitting refers to the model becoming too highly representative of the training data and thus no longer representative of the overall range of data that is supposed to be modeled.
Valid XHTML :: Valid CSS: :: Powered by WikkaWiki