The article presents the MISO model, a statistical method for analyzing RNA sequencing (RNA-seq) data to identify isoform regulation. The model estimates the expression of alternatively spliced exons and isoforms and assesses confidence in these estimates. It incorporates mRNA fragment length distribution in paired-end RNA-seq to improve the estimation of alternative splicing levels. MISO also detects differentially regulated exons or isoforms. The study applies MISO to analyze the role of the RNA splicing factor hnRNP H in regulating alternative cleavage and polyadenylation, supported by UV cross-linking–immunoprecipitation sequencing (CLIP-seq) analysis in human cells. The results provide a probabilistic framework for RNA-seq analysis, offer functional insights into pre-mRNA processing, and yield guidelines for the optimal design of RNA-seq experiments for studies of gene and isoform expression. Alternative splicing is crucial for gene function and is involved in development, differentiation, and disease. RNA-seq has become a powerful tool for characterizing the transcriptome, revealing that most human genes are alternatively spliced. MISO improves the accuracy of isoform quantification by considering both exon and isoform abundance, as well as the length distribution of library inserts in paired-end RNA-seq. The model provides confidence intervals for isoform abundance estimates and detects differential expression. MISO was validated against quantitative reverse-transcription PCR (qRT-PCR) data, showing strong correlation between MISO estimates and qRT-PCR measurements. The study also demonstrates that MISO can detect differentially expressed isoforms and validate their regulation through CLIP-seq analysis. The results suggest that hnRNP H plays a role in regulating alternative polyadenylation, with increased CLIP tag density upstream of the core poly(A) site indicating its involvement in core poly(A) site usage. The study also discusses the importance of paired-end reads and insert length in RNA-seq experiments. Paired-end sequencing provides more information about alternative exons and isoforms compared to single-end sequencing. The length distribution of inserts influences the confidence with which read pairs can be assigned to isoforms. The optimal insert size for RNA-seq experiments depends on the importance of detecting splicing changes relative to efficiently capturing short mRNAs. The study shows that using longer insert lengths can increase the yield of information about splicing by threefold or more at a given sequencing depth. However, changes in mRNA fragment size can affect the priming and reverse-transcription steps and the sampling of mRNAs of different lengths. The study also highlights the importance of similar insert lengths in RNA-seq libraries for accurate comparisons of expression and splicing. Overall, the MISO model provides a detailed probabilistic framework for RNA-seq analysis, enabling more accurate and comprehensive analysis of alternative splicing at the exon or isoform level.The article presents the MISO model, a statistical method for analyzing RNA sequencing (RNA-seq) data to identify isoform regulation. The model estimates the expression of alternatively spliced exons and isoforms and assesses confidence in these estimates. It incorporates mRNA fragment length distribution in paired-end RNA-seq to improve the estimation of alternative splicing levels. MISO also detects differentially regulated exons or isoforms. The study applies MISO to analyze the role of the RNA splicing factor hnRNP H in regulating alternative cleavage and polyadenylation, supported by UV cross-linking–immunoprecipitation sequencing (CLIP-seq) analysis in human cells. The results provide a probabilistic framework for RNA-seq analysis, offer functional insights into pre-mRNA processing, and yield guidelines for the optimal design of RNA-seq experiments for studies of gene and isoform expression. Alternative splicing is crucial for gene function and is involved in development, differentiation, and disease. RNA-seq has become a powerful tool for characterizing the transcriptome, revealing that most human genes are alternatively spliced. MISO improves the accuracy of isoform quantification by considering both exon and isoform abundance, as well as the length distribution of library inserts in paired-end RNA-seq. The model provides confidence intervals for isoform abundance estimates and detects differential expression. MISO was validated against quantitative reverse-transcription PCR (qRT-PCR) data, showing strong correlation between MISO estimates and qRT-PCR measurements. The study also demonstrates that MISO can detect differentially expressed isoforms and validate their regulation through CLIP-seq analysis. The results suggest that hnRNP H plays a role in regulating alternative polyadenylation, with increased CLIP tag density upstream of the core poly(A) site indicating its involvement in core poly(A) site usage. The study also discusses the importance of paired-end reads and insert length in RNA-seq experiments. Paired-end sequencing provides more information about alternative exons and isoforms compared to single-end sequencing. The length distribution of inserts influences the confidence with which read pairs can be assigned to isoforms. The optimal insert size for RNA-seq experiments depends on the importance of detecting splicing changes relative to efficiently capturing short mRNAs. The study shows that using longer insert lengths can increase the yield of information about splicing by threefold or more at a given sequencing depth. However, changes in mRNA fragment size can affect the priming and reverse-transcription steps and the sampling of mRNAs of different lengths. The study also highlights the importance of similar insert lengths in RNA-seq libraries for accurate comparisons of expression and splicing. Overall, the MISO model provides a detailed probabilistic framework for RNA-seq analysis, enabling more accurate and comprehensive analysis of alternative splicing at the exon or isoform level.