Services
Promoter - 2.0
Transcription start sites in vertebrate DNA
 |
Promoter2.0 predicts transcription start sites of vertebrate PolII promoters in DNA sequences. It has been developed as an evolution of simulated transcription factors that interact with sequences in promoter regions. i It builds on principles that are common to neural networks and genetic algorithms.
Submission
Sequence submission: paste the sequence(s) and/or upload a local file
Restrictions
At most 50 sequences and 1,500,000 nucleotides in total per submission.
Confidentiality
The sequences are kept confidential and will be deleted after processing.
CITATIONS
For publication of results, please cite:
Promoter 2.0: for the recognition of PolII promoter sequences.
Steen Knudsen
Bioinformatics 15, 356-361, 1999.
Output format
DESCRIPTION
For each input sequence the name and length are first printed, followed by a table in the form:
Position
where 'Position' is a position in the sequence, 'Score' is the prediction score for a transcription start site occurring within 100 base pairs upstream from that position and 'Likelihood' is a descriptive label associated with that score. The scores are always positive numbers; they are labelled as follows:
| below 0.5 |
ignored
|
0.5 - 0.8 |
Marginal prediction
|
0.8 - 1.0 |
Medium likely prediction
|
above 1.0 |
Highly likely prediction
| |
Consult the performance notes for comments on the prediction scores.
The input sequence will be included in the output, preceeding the predictions
if "Full output" has been selected.
EXAMPLE OUTPUT
Promoter 2.0 Prediction Results
INPUT SEQUENCE:
>gi_209811_gb_J01917_ADRCG Adenovirus type 2, complete genome.
CATCATCATAATATACCTTATTTTGGATTGAAGCCAATATGATAATGAGGGGGTGGAGTT
TGTGACGTGGCGCGGGGCGTGGGAACGGGGCGGGTGACGTAGTAGTGTGGCGGAAGTGTG
ATGTTGCAAGTGTGGCGGAACACATGTAAGCGCCGGATGTGGTAAAAGTGACGTTTTTGG
TGTGCGCCGGTGTATACGGGAAGTGACAATTTTCGCGCGGTTTTAGGCGGATGTTGTAGT
AAATTTGGGCGTAACCAAGTAATGTTTGGCCATTTTCGCGGGAAAACTGAATAAGAGGAA
GTGAAATCTGAATAATTCTGTGTTACTCATAGCGCGTAATATTTGTCTAGGGCCGCGGGG
ACTTTGACCGTTTACGTGGAGACTCGCCCAGGTGTTTTTCTCAGGTGTTTTCCGCGTTCC
GGGTCAAAGTTGGCGTTTTATTATTATAGTCAGCTGACGCGCAGTGTATTTATACCCGGT
GAGTTCCTCAAGAGGCCACTCTTGAGTGCCAGCGAGTAGAGTTTTCTCCTCCGAGCCGCT
CCGACACCGGGACTGAAAATGAGACATATTATCTGCCACGGAGGTGTTATTACCGAAGAA
ATGGCCGCCAGTCTTTTGGACCAGCTGATCGAAGAGGTACTGGCTGATAATCTTCCACCT
CCTAGCCATTTTGAACCACCTACCCTTCACGAACTGTATGATTTAGACGTGACGGCCCCC
GAAGATCCCAACGAGGAGGCGGTTTCGCAGATTTTTCCCGAGTCTGTAATGTTGGCGGTG
CAGGAAGGGATTGACTTATTCACTTTTCCGCCGGCGCCCGGTTCTCCGGAGCCGCCTCAC
CTTTCCCGGCAGCCCGAGCAGCCGGAGCAGAGAGCCTTGGGTCCGGTTTCTATGCCAAAC
CTTGTGCCGGAGGTGATCGATCTTACCTGCCACGAGGCTGGCTTTCCACCCAGTGACGAC
GAGGATGAAGAGGGTGAGGAGTTTGTGTTAGATTATGTGGAGCACCCCGGGCACGGTTGC
AGGTCTTGTCATTATCACCGGAGGAATACGGGGGACCCAGATATTATGTGTTCGCTTTGC
TATATGAGGACCTGTGGCATGTTTGTCTACAGTAAGTGAAAATTATGGGCAGTCGGTGAT
AGAGTGGTGGGTTTGGTGTGGTAATTTTTTTTTAATTTTTACAGTTTTGTGGTTTAAAGA
PREDICTED TRANSCRIPTION START SITES:
gi_209811_gb_J01917_ADRCG Adenovirus type 2, complete genome., 1200 nucleotides
Position Score Likelihood
600 1.063 Highly likely prediction
Performance notes
The accuracy of the software has been tested on a set of 100 vertebrate promoters. The positions scoring 0.5-0.8 (Marginal predictions) contain about 65% true transcription start sites within 100 base pairs upstream. The positions scoring 0.8-1.0 (Medium likely predictions) are about 80% true. Finally, the positions scoring above 1.0 (Highly likely predictions) are about 95% true. On average, the software picks up about 80% of all PolII promoters. These numbers are rough estimates based on a limited test set.
For a favorable comparison of this software to other promoter prediction software, see:
Eukaryotic promoter recognition.
J.W. Fickett and A.G. Hatzigeorgiou.
Genome Res. 7(9), 861-878, 1997.
References
, Bioinformatics,15, 356-361,1999
Abstract
Motivation: a new approach to the prediction of eukaryotic Pol II
promoters from DNA sequence takes advantage of a combination of elements
similar to neural networks and genetic algorithms to recognize a set of
discrete subpatterns with variable separation as one pattern, a promoter. The
neural networks use as input a small window of DNA sequence, as well as the
output of other neural networks. Through the use of genetic algorithms, the
weights in the neural networks are optimized to maximally discriminate between
promoters and non-promoters.
Results: after several thousand generations of optimization, the
algorithm was able to discriminate between vertebrate promoter and non-promoter
sequences in a test set with a correlation coefficient of 0.63. In addition,
all five known transcription start sites on the plus strand of the complete
Adenovirus genome were within 161 bp of 35 predicted transcription start
sites. On standardized test sets consisting of human genomic DNA, the
performance of Promoter 2.0 compares well with other software developed for the
same purpose.