Skip to content

Visualization

Spl-IsoQuant provides a visualization tool to help interpret and explore the output data. The goal of this visualization is to create informative plots that represent transcript usage and splicing patterns for genes of interest. Additionally, we provide global transcript and read assignment statistics from the Spl-IsoQuant analysis.

Running the visualization tool

To run the visualization tool, use the following command:

python splisoquant_visualize.py <output_directory> --gene_list <gene_list> [options]

Command line options

  • output_directory (required): Directory containing Spl-IsoQuant output files.
  • --gene_list (required): Path to a .txt file containing a list of genes, each on its own line.
  • --viz_output: Optional directory to save visualization output files. Defaults to the main output directory if not specified.
  • --gtf: Optional path to a GTF file if it cannot be extracted from the Spl-IsoQuant log.
  • --counts: Use counts instead of TPM files for visualization.
  • --ref_only: Use only reference transcript quantification instead of transcript model quantification.
  • --filter_transcripts: Filter transcripts by minimum value occurring in at least one condition.
  • --find_genes [N]: Rank genes by a combined rank of expression, expression difference and transcript usage deviance relative to the wild_type condition, and visualize the top N genes (default is 100).
  • --known_genes_path: Path to a file with known target genes (one per line), used together with --find_genes to mark known targets in the ranking.
  • --read_group_strategy: Grouping strategy whose grouped counts are visualized, named as it appears in the grouped counts file names (SAMPLE_ID.gene_grouped_<strategy>_counts.tsv), e.g. file_name, barcode, tag_CB for --read_group tag:CB or file0_col1 for a grouping table. Defaults to the first strategy given with --read_group in the original run that has grouped counts in the output directory; strategies Spl-IsoQuant adds on its own (e.g. barcode in single-cell modes, file_name for several input files) are not recorded, so for them the alphabetically first strategy is used. An unknown strategy is reported as an error.

Input files

Read assignment statistics are taken from SAMPLE_ID.read_info.tsv.gz, or from the deprecated SAMPLE_ID.read_assignments.tsv.gz when that is the only per-read file present. If neither is in the output directory, re-run Spl-IsoQuant with --large_output read_info.

Grouped counts are plotted per condition only when the number of groups is small (up to 100, as for --read_group file_name). Single-cell and spatial runs produce one group per barcode or spot, so the visualizer falls back to the ungrouped counts for them and reports this on the console.

Output

The visualization tool generates the following plots based on the Spl-IsoQuant output:

  1. Transcript usage profiles: For each gene specified in the gene list, a plot showing the relative usage of different transcripts across conditions or samples.

  2. Gene-specific transcript maps: Visual representation of the different splicing patterns of transcripts for each gene, allowing easy comparison of exon usage and alternative splicing events.

  3. Global read assignment consistency: A summary plot showing the overall consistency of read assignments across all genes and transcripts analyzed.

  4. Global transcript alignment classifications: A chart or plot representing the distribution of different transcript alignment categories (e.g., full splice match, incomplete splice match, novel isoforms) across the entire dataset.

These visualizations provide valuable insights into transcript diversity, splicing patterns, and the overall quality of the Spl-IsoQuant analysis.