The Genomic Threading Database Help Page

The Genomic Threading Database Help Page

To print this page just select File -> Print, at the top of this window.
For any other queries email: Dr D.Buchan

By clicking on the relevant links on the main page you may:

Carry out a key word search on the database

View a summary of the coverage of predictions

Download a GTD list containing the top prediction for each sequence in a given genome.

The following is a description of how to perform key word searches using the seach form and how to interpret the results pages:

Select Genomes

Select which genomes you wish to search. The sources and version numbers of the sequences can now be found on the summary page.

Multiple genomes may be selected by holding down the Shift key and dragging down the list.

Alternatively, hold down the Ctrl key and click on each entry.

Search Options

Type keywords, phrases, or codes into the boxes provided.

DO NOT USE QUOTATION MARKS SUCH AS " OR '
SEARCHES ARE CASE SENSITIVE
DO NOT USE WILDCARDS SUCH AS *, ?, % or & (SEE BELOW)

Select which field you wish to search:

Gene description or ID - the description lines from each sequence entry.
PDB description - the headers of the Protein Data Bank files in the current fold library.
SCOP description - Structural Classification of Proteins description of folds in the current library.
PDB code - the 4 character Protein Data Bank ID of folds in the current library.
SCOP code - the Structural Classification of Proteins code of folds in the current library.

Full keywords/phrases/codes need not be entered. For example, you may search for 'All alpha, Globin-like proteins' either by entering this phrase and selecting 'SCOP description', or by simply entering the keyword 'a.1.' and selecting 'SCOP code'. Entering the term 'oxidase' and selecting 'PDB description' will also return results for 'peroxidase'. In order to search for the term 'oxidase' only, you should add spaces before and after the search term such as ' oxidase ' (without the quotes).

If you wish to search by one key word only, you must select '- None -' next to the other text box.

Output Options

You may limit the number of hits displayed per chromosome per genome (N.B. hits are initially ranked by p-value).

Please note the speed at which the results table is drawn is dependent on the number of hits you select as well as the speed of your connection.

You may also select how you wish to re-rank the results table:

P-value - a measure of significance of each hit. Lower p-values indicate more confident hits.
Score - the raw output score form the neural network. Higher scores indicate more confident hits.
Pair Energy - traditional threading potential - pairwise potential of mean force, used to evaluate the alignment between target and template. Lower energy generally indicates a better alignment.
Solvation Energy - traditional threading potential - relating to the degree of residue burial, used to evaluate the alignment between target and template. Lower energy generally indicates a better alignment.
Alignment score - the score of the sequence alignment between the target and template. Higher scores generally indicate closer relationships.

Check the Machine readable checkbox for machine readable output, obviously.

Keyword Search Results Page

The Keyword search results page displays data on the top hits found as a single table.

The table is divided up into sub tables showing hits per chromosome, and each of these is divided into several columns.

The Confidence column gives an indication of the strength of the hit, ranging from 'certain' to 'guess'. Each cell is also colour coded for ease of use.

Certain - p-value < 0.0001
High - p-value < 0.001
Medium - p-value < 0.01
Low - p-value < 0.1
Guess - p-value >= 0.1

To display an alignment, click on the PDB code within a cell in the Alignment column. The alignment will be displayed in a popup window. To print the alignment select File -> Print, at the top of the window.

To obtain/display the model generated from the alignment click the link which says "CLICK HERE TO DOWNLOAD MODEL IN PDB FORMAT" in the popup window. You can configure your browser to associate PDB files with a molecular graphics viewer such as RasMol.

To view the full SCOP entry information, click on the link within a cell in the SCOP code column.

To view the full PDB entry information, click on the image within a cell in the Structure column.

BLAST Search Results Page

The BLAST search results page shows the top sequence matches within the GTD to your target sequence.

If a close match to your target sequence is found click on the MD5 sequence identifier link to obtain the top hits for that sequence.

Clicking the MD5 sequence identifier link will take you to the Keyword Search Results Page (Clicking the MD5 sequence identifier link allows you to automatically perform a keyword search for the MD5 sequence indentifier on the given genome).