Computational protocol: Enhanced transcriptome maps from multiple mouse tissues reveal evolutionary constraint in gene expression

Similar protocols

Protocol publication

[…] RNA-seq reads were aligned to the human (hg19) and mouse (mm9) genomes using the STAR 1.9 software with up to ten mismatches per paired alignment and without using the annotations. Only alignments for reads mapping to ten or fewer loci were reported. Mapped reads were used to generate contigs, splice junctions and de novo transcript models. Non-parametric irreproducibility ascertainment (npIDR) was used for genomic elements (such as splice junctions, exons, transcripts, etc) by requiring npIDR≤0.1. Strand-specific contigs were called from merged biological replicates independently of the annotations and assessed with npIDR. Cufflinks 1.0.3 (ref. ) was used to assemble de novo transcripts from STAR alignments considering only uniquely mapping non-duplicated alignments crossing GU/AG junctions. The alignments from the two bio-replicates were merged before Cufflinks assembly. Cufflinks transcript models passing the npIDR threshold (npIDR≤0.1) were pooled across all experiments and further filtered by requiring CAGE data support in FANTOM study. The Cufflinks gene, transcript and exon RPKM were quantified using Flux Capacitor in each bio-replicate, and the resulting RPKM were assessed for reproducibility using npIDR. AVG read density was computed by using only uniquely mapped reads, separately for each bio-replicate and for each strand without applying npIDR. The quantitative assessment of splicing at the level of splice junctions was done by using intron-centric metrics. More details on RNA-seq data processing are in . […]

Pipeline specifications

Software tools STAR, Cufflinks
Application RNA-seq analysis
Diseases Neoplasms