A novel parameter describing the influence of a mutation on alternative splicing.

To calculate the "Splicing Effect" (SE value), we need only consider one mRNA isoform. We choose the longer isoform. This allows us to handle mutations causing either whole exon skipping or use of alternative splice sites with the same equation.

We then define F_m as the frequency of our isoform with the mutation and F_w as the frequency of our isoform without the mutation (that is, in the wild-type exon sequence).

The SE value is then calculated as follows:

The SE value can then be shown to reside within a range of [-1, +1].

For example, for a mutation causing 100% skipping of the constitutive wildtype exon, F_m = 0 and F_w = 100. Thus, SE = (0 − 100) / (0 + 100) = -1.

Correspondingly, a mutation causing 100% retention of what would otherwise be considered an intron has the most positive SE value. In this case, F_m = 100 and F_w = 0. Thus, SE = (100 − 0) / (100 + 0) = +1.

As an additional example, consider a situation in which the longer isoform is shifted from 50% in wild-type to 75% with mutation. Then, F_m = 75 and F_w = 50. Thus, SE = (75 − 50) / (75 + 50) = +0.2.

As a final example, consider a situation in which the longer isoform is shifted from 50% in wild-type to 25% with mutation. Then, F_m = 25 and F_w = 50. Thus, SE = (25 − 50) / (25 + 50) = -0.33.

In order to compare and synthesize data from multiple studies using various experimental methods, each splicing effect is assigned an accuracy level. This level is determined based on the published description of the methods used to measure the relative quantities of alternatively spliced isoforms in the mutant and wild-type sequences. We have established a four-level system using the following definitions:

• Level 0 (low) - when no experimental figure or details are provided for the mutation
• Level 1 (moderate) - when non-quantitative PCR was only used for describing the appearance of a new isoform due to the mutation; In this case we estimate the wildtype and mutant isoform frequencies ourselves based on the published figures. These rough estimations are counted only in quarters (0%, 25%, 50%, 75%, 100%), thirds (0%, 33%, 67%, 100%), or minor/major isoform appearance (10%, 90%).
• Level 2 (good) - when semi-quantitative PCR or another similar experimental technique was used for characterization of the wildtype and mutant isoform frequencies
• Level 3 (high) - when quantitative PCR was used for characterization of the wildtype and mutant isoform frequencies

A sequential, unique, internal identifier for each mutation in the ASMD (for example: "21asmd")

The ASMD IDs are intended to be perpetual, so they cannot be reused/recycled for a new mutation; They can, however, be expired for obsolete entries (for example, an entry could be expired if a new publication shows that this particular mutation is irrelevant to splicing). The numbering is sequential and does not necessarily reflect any ordering based on the properties of the mutations.

Exon numbering in genes with alternative splicing can be ambiguous because the same exon can have a different consecutive number in different isoforms. Our exon numbers may therefore differ from those in the literature (e.g. exon #18 in the referenced article may be exon #14 in our database). [NOTE: This can also be due to the earliest publications not taking into account 5′-untranslated exons.]

Therefore, while presenting an exon number, we simultaneously provide a link to the gene isoform from which this number derives. All gene isoforms are taken from the Exon-Intron Database (Shepelev and Fedorov 2006, human genome release hs36p1.dEID).

NOTE: All gene information, including exon number, is taken from the Exon-Intron Database (EID), version hs36p1. These exon numbers may not agree with those in other references. See paper: Shepelev V., Fedorov A. Advances in the Exon-Intron Database (EID). Briefings in Bioinformatics 2006, 7: 178-185. (PDF 111KB)

Denotes the position of the mutation (in nucleotides) from the 5′-end of the exon containing it

For instance, position "@1" means that this is the first nucleotide of the exon. For mutations longer than one nucleotide (for example, a 3-nt deletion), the position reflects the first mutant nucleotide from the 5′-end. For ambiguous positions (for example, insertion of "A" inside an "AAAAA" segment or deletion of "AG" from an "AGAG" segment) the position reflects the uppermost 5′-position among all possible variants.

Types of alternative splicing caused by mutations.

NOTE: one mutation can cause several splicing effects (for example, simple skipping and alternative acceptor site). Therefore, the total number of splicing effects in the ASMD is greater than the total number of mutations.

Currently the ASMD allows for the following 6 splicing effect types:

• Alternate acceptor site - activation of a cryptic splice acceptor site (3′ of intron / 5′ of exon); also known as "alternate 3′-splicing", according to Jane Wu (ref).
• Alternate donor site - activation of a cryptic splice donor site (5′ of intron / 3′ of exon); also known as "alternate 5′-splicing", according to Jane Wu (ref).
• Simple skipping - skipping of a single exon containing the mutation.
• Cassette skipping - skipping of several adjacent exons, one of which contains the mutation. For example, cassette skipping (3-5) means skipping of exons 3, 4, and 5 where the mutation resides in one of those exons, such as exon 3.
• Trans skipping - skipping of a single exon NOT containing the mutation. For example, trans skipping (2) means skipping of exon 2 where the mutation is located in another exon, such as exon 1.
• Trans cassette skipping - skipping of several adjacent exons, none of which contain the mutation. For example, trans cassette skipping (3-5) means skipping of exons 3, 4, and 5 where the mutation is located in another exon, such as exon 2.