1. What is the Human Transcript Database ?

2. How do I use the database ?

3. How is this database different from Unigene ?

4. How is the database made ?

5. Credits

Characteristic Number:

There are several databases exist that cluster all the information relevant to potential human transcripts into a single location. The sources of this data include complete human transcript information as well as EST data and theoretical gene predictions. Due to the problems associated with EST and gene-prediction data these databases contain many non-biological transcripts. Furthermore, when consensus sequences are created they are electronic and no physical clone exists that can be obtained by an interested investigator. To eliminate the noise and provide a clone based database we have assembled the Human Transcript Database. This database consists solely of sequences that are accurate and derived from a single clone which may be obtained by any researcher.

The Human Transcript Database (HTD) can be accessed through the search engine. This web interface allows one to perform a number of searches of the database and will return a table displaying all of the matching sequences that are present in the database. There are two basic search mechanisms - content based and sequence based.

To search the database for the presence of keywords - for instance a particular author, use the search form. This searches the database for any keyword or characteristic of the entry. To search for sequences based upon their length, enter the minimum size desired or the size range desired and press the search key. A table will be displayed that shows all of the matching sequences. The length search may be performed in combination with the keyword search.

The database can also be used to compare a given sequence to all sequences in the HTD. This can be accomplished by either uploading a sequence to the web page or pasting a sequence into the given space. Note that there should be no header sequences in the pasted sequence but an uploaded sequence should be in FASTA format (first line is a description). The input sequence is compared to the database by a blast program (Reference: Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schaffer, Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman (1997), 'Gapped BLAST and PSI-BLAST: a new generation of protein database search programs', Nucleic Acids Res. 25:3389-3402 Abstract). The output is displayed as a graph with an associated table.

For more sensitive searches, the cutoff value for blast searches can be set (bottom of form) - please note that the blast program is very good at returning nearly exact matches but inexact matches are both less reliable and more computationally intensive.

The table that is displayed shows the data contained in a number of fields and links to other databases. The first column contains a link to the GenBank entry that is maintained at the National Center for Biotechnology Information. This represents the original source of data that the Human Transcript Database was built from. The second field contains a link to the unigene clustering database. This database clusters together all similar sequences including EST sequences and predicted sequences. The third column contains the accession number for the entry - this is a unique identifier that is often used for referencing this sequence. The fourth column contains the GID number which is also a unique number for each sequence but may change over time. The fifth column contains a brief description of the entry. The sixth column represents the length of the clone in the database. The seventh column contains a link to the databank entry which will display all of the information we maintain on the sequence. For sequence comparison searches, the score of the match is displayed in a eighth column.

Human Transcript Database assembly methods

Description of Method:
The asn1 files from the NCBI were downloaded and examined to find the sequences that claimed to be mRNA sequences or were of unknown type. The accession numbers from these sequences were then used to parse through the corresponding GenBank file format and extract all of the information present in the database. At this point, the sequences were tested for the presence of one of a number of keywords (e.g. BAC). Entries containing one of these keywords were removed from the database. Following this culling step, the final list of all of the accession numbers was used to parse through the fasta file of all of GenBank to create a concatenated fasta file of all of the transcripts to be used for Blast searches. Lastly, the database was examined for transcripts that had been cut up into individual exons and the joined transcript was produced.

Caveats:
The approach we have taken focuses on the incorporation of as many transcripts as possible. Reasonable effort has been made to exclude those sequences that are not derived from mRNAs, however such sequences exist within the database. Chimeric and genomic cDNA clones have not been rigorously excluded from the database.

Disclaimer:
The database assembled here is intended to assist in the acquisition of scientific knowledge. No claims as to the completeness or accuracy of the database is made.