IsoQuant 3.12.2

Fix package naming.

IsoQuant 3.12.1

Add pip installation.

IsoQuant 3.12.0

• TSV output files do not have a # in headers for easier downstream processing (thanks to @yinshiyi).

• Fix Python 3.14 compatibility.

• Add --barcoded_bam flag for reading barcodes and UMIs directly from a input BAM.

• Add --barcode2barcode option for Visium HD data processing.

• Dramatic speed-up for 10x and universal barcode calling on large whitelists.

• Per-barcode counts grouping is now performed automatically if the barcodes are set.

IsoQuant 3.11.1

• Fix Python 3.8

• Fix MTX conversion.

IsoQuant 3.11.0

• --read_group now supports multiple read grouping strategies. You can now simultaneously group counts by samples, BAM tags, read attributes provided in separate TSV files or within read ids themselves.

• New --large_output option to control which large output files are generated.

• Significant performance optimizations.

• Experimental release of the single-cell/spatial IsoQuant pipeline. Official release will follow soon.

IsoQuant 3.10.0

• New --unmapped_bam option for providing unmapped BAM files typical for PacBio CCS data.

• New --polya_trimmed option to indicate polyA-trimmed reads (thanks @hmutpw for the suggestion #342).

• New --process_only_chr option to process a specific list of chromosomes.

IsoQuant 3.9.0

• Secondary alignments are not used by default from now on. It significantly improves running time and RAM consumption, but barely affects the results' quality.
  Use --use_secondary to process secondary alignments.

• New options that force IsoQuant to use only a faction of reads in high-coverage loci.
  Significantly improves running time and RAM consumption, but affects gene/isoform counts.
  New default behavior only affects small chromosomes and scaffolds (<500kbp).

  In some cases, high-coverage regions take too much time to process due to extreme number of mapped reads,
  especially chrM (up to 10x longer compared to normal chromosomes). However, using only a fraction of these
  reads is enough to obtain reliable results.

  These options allow to process only up to given number of reads mapping to a high-coverage loci on short and normal chromosomes:

  • --max_coverage_small_chr (default value is 1 million);
  • --max_coverage_normal_chr (default value is infinity, so usual chromosomes are not affected by default even if some genes have extreme coverage).

• New option --discard_chr to discard a list chromosomes from the analysis.

IsoQuant 3.8.0

• Fixed --report_canonical preset (#332, thanks to @wwliao).
• Fixed counts for novel genes in discovered_gene_counts.tsv and discovered_gene_tpm.tsv (#337, thanks to @yjliuhub).
• Fixed --genedb_output option (#335, thanks to @YalanBi).

IsoQuant 3.7.1

• Support for indexing BAMs with large chromosomes, fixes #327. Thanks to @maol-corteva.
• CDS features are now used when exon features are absent, fixes #309.
• Chromosome names are now checked for consistency between reference genome, annotation and BAM file provided. Only overlapping chromosome names are used if inconsistent.
• Fixed SQANTI-like output headers, fixes #318.
• Some minor cosmetics.

IsoQuant 3.7.0

• Optimized grouped counts output. By default, all counts are stored in linear format, which saves time and disk space.
  Matrices with small number of columns are automatically converted to usual matrix in TSV format,
  larger matrices typical for single-cell and spatial data are converted to MTX format.
  See --counts_format paramter for options. It is also possible to convert counts after IsoQuant is finished using src/convert_grouped_counts.py.
  Fixes issues mentioned in #248

• Renamed counts related to discovered transcripts and genes to avoid confusion.

• New options --indexing_options and --mapping_options that allow to pass options to the indexing and mapping commands.
  Fixes #284 and #259

• STARlong is now an alternative options for aligning, can be set via --aligner starlong (not recommended for ONT reads).
  Fixes #284

• Exon/splice junction counts now only come from reads assigned to the same strand, fixes #253

• Use only gene-assigned reads for exon counting, fixes #283

• Fixed rare serialization bug #304