Methods and computer software for detecting splice variants

• US 20070148667A1
• Filed: 10/02/2006
• Published: 06/28/2007
• Est. Priority Date: 09/30/2005
• Status: Active Grant

• Current Assignee: Affymetrix, Inc. (Thermo Fisher Scientific Incorporated)
• Original Assignee: Affymetrix, Inc. (Thermo Fisher Scientific Incorporated)
• Inventors: Wang, Hui, Cawley, Simon, Williams, Alan, Clark, Tyson, Blume, John
• Granted Patent: US 7,962,291 B2
• Time in Patent Office: Days
• US Class Current: 435/6
• CPC Class Codes: G16B 20/20 Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
• CPC Class Codes: G16B 25/00 ICT specially adapted for hybridisation; ICT specially adapted for gene or protein expression

Abstract

Methods and software products for analysis of alternative splicing are disclosed. In general the methods involve normalizing probe set or exon intensity to an expression level measurement of the gene. The methods may be used to identify tissue-specific alternative splicing events.

Claims

1. A computer-implemented method for identifying exons that are differentially spliced between a first and second sample comprising:

(a) obtaining an exon intensity measurement for a first exon in a first gene in said first sample and in said second sample;

(b) obtaining a gene level measurement for said first gene in said first sample and in said second sample;

(c) obtaining a first normalized intensity measurement for said first exon in said first sample and a second normalized intensity measurement for said first exon in said second sample by dividing the intensity measurements obtained in (a) by the gene level measurements obtained in step (b) wherein the exon intensity in said first sample is divided by the gene level measurement in said first sample and the exon intensity in said second sample is divided by the gene level measurement in said second sample;

(d) calculating a splicing index measurement for said first exon, wherein the splicing index is equal to the log₂ of the normalized intensity for said first exon in said first sample divided by the normalized intensity for said first exon in said second sample; and

(e) identifying said first exon as being differentially spliced if the absolute value of the splicing index obtained in (d) is greater than a threshold value.

2. The method of claim 1 where the threshold value is 1-2.

3. The method of claim 1 wherein the threshold value is 2.

4. The method of claim 1 wherein the normalized intensity measurements of step (c) are obtained using the following equation:

n_i,j,k=e_i,j,k/g_i,k where e_i,j,k is the exon signal estimate of the i exon, j experiment, and k gene and g_i,k is the gene level signal estimate of the j experiment and k gene.

6. The method of claim 1 wherein the exon is identified as being overrepresented in said first sample relative to said second sample if the splicing index is greater than zero.

7. The method of claim 1 wherein the exon is identified as being overrepresented in said first sample relative to said second sample if the splicing index is greater than 2.

8. The method of claim 1 wherein the exon is identified as being underrepresented in said first sample relative to said second sample if the splicing index is less than 0.

9. The method of claim 1 wherein the exon is identified as being underrepresented in said first sample relative to said second sample if the splicing index is less than −

2.

5. The method of claim 5 wherein a first value of n_i,j,k is calculated for a first sample and a second value of n_i,j,k is calculated for a second sample and the splicing index is calculated by dividing the value of the n_i,j,k for said first sample by the n_i,j,k for said second sample to obtain a third value and calculating the log₂ of said third value.

10. A computer-implemented method for identifying exons that are differentially spliced between a first and second sample comprising:

a. obtaining an exon intensity measurement for a first exon in a first gene in said first sample and in said second sample;

b. obtaining a gene level measurement for said first gene in said first sample and in said second sample;

c. obtaining a first normalized intensity measurement for said first exon in said first sample and a second normalized intensity measurement for said first exon in said second sample using the following equation;

n_i,j,k=e_i,j,k/g_i,k where e_i,j,k is the exon signal estimate of the i exon, j experiment, and k gene and g_i,k is the gene level signal estimate of the j experiment and k gene;

d. obtaining a splicing index measurement for said first exon, wherein a first value of n_i,j,k is calculated for a first sample and a second value of n_i,j,k is calculated for a second sample and the splicing index is calculated by dividing the value of the n_i,j,k for said first sample by the n_i,j,k for said second sample to obtain a third value and calculating the log₂ of said third value; and

e. identifying said first exon as being differentially spliced if the absolute value of the splicing index obtained in (d) is greater than a threshold value.

11. The method of claim 10 wherein the threshold value is between about zero and about 2.

12. The method of claim 10 wherein the exon intensity measurements are obtained by hybridizing nucleic acid derived from the sample to an array of probes wherein there is a probeset comprising a plurality of probes for each exon to be interrogated and there are more than 10,000 different probesets.

13. A method for predicting if an exon in a gene is differentially spliced in a first and a second sample sets, comprising:

obtaining a first exon level intensity measurement for said exon in said first sample set and a second exon level intensity measurement for said exon in said second sample set;

assume under the null hypothesis of no alternative splicing for said exon that all exons of said gene are expressed at levels that are proportional to each other in said first and second sample sets;

fit a model that predicts exon response under said null hypothesis;

construct a statistic that measures how deviant the observed data is from the model;

use the statistic to construct a p-value; and

analyze the p-value to predict if the exon is differentially spliced.

14. The method of claim 13 wherein the model is e_i,j,k=α

_i,k,g_j,k where e_i,j,k is the signal of the i-th exon of the j-th sample of the k-th gene, g_j,k is the signal of k-th gene in the j-th sample, and α

_i,k is the ratio of exon i signal to its gene signal.

15. The method of claim 13 wherein the model is e_i,j,k=α

_i,kp_i,j,kg_j,k where e_i,j,k is the signal of the i-th exon of the j-th sample of the k-th gene, g_j,k, is the signal of k-th gene in the j-th sample, α

_i,k is the ratio of exon i signal to its gene signal in the sample where it is maximally expressed, and 0≦

P_i,j,k≦

1 is a Splicing Index that estimates the proportionate expression of this exon of this gene in tissue j.