Predicting Splice Variant from DNA Chip Expression Data

Table 1.

Splice Variant Prediction from Three Rat Tissues

A.
Selection ratio (R) 5-fold 7-fold 10-fold 15-fold 20-fold
Tissue type Algorithm gene probe gene probe gene probe gene probe gene probe
Heart SP 183 216 138 165 100 117 60 73 46 58
Heart SP + NB 10 29 8 21 5 13 5 12 5 12
Liver SP 55 67 43 54 22 32 13 21 9 17
Liver SP + NB 3 9 3 9 3 9 3 9 3 9
Skeletal muscle SP 126 144 85 92 48 52 26 28 19 21
Skeletal muscle SP + NB 8 18 2 4 2 4 1 2 1 2
Total SP 328 414 243 303 158 197 93 120 68 94
Total SP + NB 20 54 12 32 9 24 8 21 8 21
B.
Selection ratio (R) 5-fold 7-fold 10-fold 15-fold 20-fold
Tissue type Algorithm gene probe gene probe gene probe gene probe gene probe
3 tissues (HLS) SP 864 2216 680 1411 469 819 283 419 208 269
3 tissues (HLS) SP + NB 227 1192 133 624 69 281 35 114 17 58
  • (A) Splice variant prediction from triplicate control experiment. Total RNA was extracted from rat heart, liver, and skeletal muscle tissues. Independent RNA labeling and chip hybridization experiments were performed as triplicate for each tissue sample. Potential splice variants were predicted from each set of triplicate data using SPLICE (SP) algorithm alone or in combination with NEIGHBORHOOD (NB) algorithm. Total number of predictions from each tissue set was calculated. (B) Splice variant prediction from three different rat tissues. To generate the data set of three different tissues, the mean CSS value of each tissue triplicate was calculated and appended into the same table. Splice variant predictions were performed using the combined data set from the three tissues.

This Article

  1. Genome Res. 11: 1237-1245

Preprint Server