Software Tools :: Genefinding and Codon Analysis
Gene finding programs have used three approaches in predicting the locations of genes:
- search by content -- locating open reading frames,
regions that have G+C content and codon usage characteristic of
coding regions, etc.
- search by signal -- locating short sequence motifs
associated with genes, such as promoter and transcription factor
binding sites, or splice sites
- search by homology -- using the sequence of a gene from one
organism to identify homologs in other organisms; aligning EST
sequences to genomic sequences
Today the most accurate gene-finding methods use machine learning
techniques such as neural nets, decision trees, and hidden Markov models
to evaluate the information from all of these "traditional" approaches in
order to make a prediction. These programs are trained using a database of
known genes and may not yield accurate predictions for organisms that
weren't represented in the training set.
Back to
Genefinding and Codon Analysis
This website will look much better in a browser that supports
web standards, but it has been designed so
that it is still usable and accessible to any browser or web-enabled device.
|
