BIP-Splice

BIP-Splice

BIP-Splice is a secondary tool in the BIP toolbox. It's purpose is to take transcripts loaded and aligned using BIP-Align, cluster them together, and perform alternative splicing analysis.

Objective:
To perform alternative splicing analysis on transcripts present in a BIP-Align database. Transcripts are clustered according to overlap of exons. Each cluster is analyzed for exon variation between transcripts.
See splice analysis

Usage of BIP-Splice

The following entites are created by BIP-Splice:
Cluster
Groups of transcripts which overlap by at least one base pair in their exons. There are one or more transcripts per cluster. There can be many transcripts per chromosome, and only one cluster per transcript.

The steps are:

Clustering:

Clustering is a two-step process. First, transcripts are grouped by overlap, i.e. all transcripts which overlap by at least one base pair are considered part of the same cluster.
Overlap takes into account not only the genomic coordinates of the transcripts, but also the orientation, or strand. For instance, if two transcripts overlap based on genomic position, but one maps to the positive strand and the other to the negative, they are members of two different clusters.

The second step in clustering uses each transcript's exon/intron structure to refine the clusters. To be a member of a cluster, a transcript must have at least one exon which overlaps by a minimum of one base pair with the exon of another transcript. If it does not meet this criteria, then the transcript in question creates a new cluster.

Quality filtering:

The quality filter provides a parameter which requires that all clusters have a minimum number of transcripts as members. The default value is 3, though a BIP-Splice database can be created which allows singleton and doubleton clusters if desired.

Splice analysis:

Each cluster is analyzed to determine if they exhibit any alternative splicing. The alternative splicing events that are recorded are:
  • Length variation. This refers to internal splice sites of exons, and if they differ between member transcripts. The 5' and 3' ends of the transcripts are not evaluated for splice variation because of the possibility of a truncated sequence.
  • Initial casette exons. This type of exon is missing in one or more transcript. An initial exon is the 5' exon of a transcript. To be flagged as an initial cassette exon, the exon cannot occur as an internal exon in any transcript.
  • Terminal casette exons. Same as initial cassette exon, except it occurs at the 3' end.
  • Internal cassette exons. These are cassette exons which are present as an internal exon in at least one transcript of the cluster. Internal cassette exons are presumed to be the most biologically relavent because truncated sequences may create artificial occurences of initital and terminal cassette exons.

Any cluster which has at least one form of splice variation is flagged as 'variant'.
Once the above steps are complete, the database is updated with clustering and splice information.

Table information:

BIP-Splice adds new tables to the schema from BIP-Align:

Cluster table:
  • chromosome_id
  • orientation
  • chr_beg
  • chr_end
  • variant
Genomic exon table:

This table warrants a bit of explanation. During the splice analysis step, all overlapping exons in one position of a cluster are grouped together into a consensus exon. This is called the genomic exon by BIP-Splice, and this is the table which holds the informtion regarding this consensus, or genomic exon.
  • cluster_exon
  • cluster_id
  • chr_beg
  • chr_end
  • chr_length
  • variant
  • chr_sequence
  • cassette
  • cass_initial
  • cass_internal
  • cass_terminal
  • variation_5end
  • variation_3end
The clone and clone exon tables are expanded and updated by BIP-Splice as well

Clone table:

Added field: cluster_id

Clone Exon Table:

Added fields:
  • cluster_id
  • cluster_exon
  • cassette
  • cass_initial
  • cass_internal
  • cass_terminal
  • variation_5end
  • variation_3end

Table relationships:

There is a one to many relationship from the cluster table to clone table. I.e. each clone can belong to zero or one clusters only.
From the genomic exon table to clone exon table, there is a one to many relationship. From the genomic exon table to cluster, there is a many to one relationship.

BIP-Splice Database Schema

For any comments or suggestions, send an e-mail at: BIP_FEEDBACK@asu.edu
Back to: Project Overview
BIP-Splice Use Cases
BIP-Splice General Use Case
BIP-Splice Schema

Last updated: 12/12/2006
Home | Site Map | Project | Users | Participants | Sponsors | Links | Papers | Contact | FAQ | Glossary ©2006 BIP