Figure 5

Using Sweepster to find parameter settings that are good for assembling isoforms of tumor suppressor genes. Isoforms are being assembled from mapped RNA-seq reads of bladder cancer cell line T24 (SRA accession no. SRX148575). An enhanced Galaxy tool form for Cufflinks (left panel), a transcript assembly tool [13], is used to create a partial tree of the tool’s parameter space (middle panel). The tree shows the parameter space for two parameters, minimum_isoform_fraction and pre-mRNA_fraction. Minimum, maximum, and number of samples are used to customize the tree. Clicking on the interior tree node where minimum_isoform_fraction equals 0.1 launches jobs for Cufflinks that systematically sample pre-mRNA_fraction from 0 to 0.2; one job is run for each value of pre-mRNA. Each track shows transcript assemblies for tumor suppressor genes TP53 (chr17:7569899-7591385) and VHL (chr3:10182352-10196520) produced by Cufflinks for a particular set of parameters (right panel). TP53 isoforms are similar across the assemblies, but VHL isoforms are substantially different. Sweepster’s tracks can be used to visually compare assemblies produced from different parameter settings and choose the settings that yield the best assembly. Once good settings are found, Cufflinks can be run on the complete dataset using the track’s controls.