130 likes | 236 Views
Machine Learning group meeting. Yiming Zhang 02/03/2012. Preliminary results. Number of introns predicted in different chromosomes and under different criteria. Preliminary results. Expected and observed percentage of introns in different chromosomes. Preliminary results.
E N D
Machine Learning group meeting Yiming Zhang 02/03/2012
Preliminary results Number of introns predicted in different chromosomes and under different criteria
Preliminary results Expected and observed percentage of introns in different chromosomes
Preliminary results • For intron retention, Chi squared equals 8.092 with 4 degrees of freedom. The two-tailed P value equals 0.0883. • For constitutive intron, Chi squared equals 4.294 with 4 degrees of freedom. The two-tailed P value equals 0.3677
Preliminary results Number of genes which is predicted to have all introns constitutively or alternatively spliced (at least 2 introns per gene) under different criteria
Dataset and model • Rice Genome Annotation Project (Michigan State Univ.) genome sequence and gff3 annotation version 7.0. • After removing redundancy and the introns or their flanking exons which are shorter than 20 nt, 196,290 introns and 44,037 genes have been used. • OS specific model (trained by RF with 200 trees) has been used to make prediction.
Preliminary results Number of introns predicted under different criteria and number of different kinds of introns per gene
Preliminary results Number of introns predicted in different gene region and under different criteria
Preliminary results Number of introns predicted in different chromosomes and under different criteria
Preliminary results Number of introns predicted in different chromosomes and under different criteria
Preliminary results Expected and observed percentage of introns in different chromosomes
Preliminary results • For intron retention, Chi squared equals 584.436 with 11 degrees of freedom. The two-tailed P value is less than 0.0001. • For Constitutive intron, Chi squared equals 443.521 with 11 degrees of freedom. The two-tailed P value is less than 0.0001
Preliminary results Number of genes which is predicted to have all introns constitutively or alternatively spliced (at least 2 introns per gene) under different criteria