Skip to content

Releases: nanoporetech/megalodon

Squirt

02 Feb 00:41
Compare
Choose a tag to compare

This a point release including new features:

  • Allow supplementary alignments with --allow-supplementary-alignments flag (resolves #71)
  • Patch for guppy connection issues (full fix coming with future Guppy release)
  • Better Guppy warnings and help pointers
  • Add megalodon_extras calibrate generate_modified_base_stats command
  • Lower aggregation batch size default (resolves #76)
    Bug Fixes:
  • describe_alphabet error (resolves #87)
  • atomize_variants error (resolves #82)

Ray

02 Dec 00:43
Compare
Choose a tag to compare
Ray

Add calibration files for updated 5hmC model as well as megalodon_extras commands to assist in model training.

Fish School

26 Nov 23:22
Compare
Choose a tag to compare

Fix paths to 5mC model calibration files.

Peach

25 Nov 06:36
Compare
Choose a tag to compare

This a point release including new features:

  • Add 5mC Rerio Model calibration files
  • Add mapping trimming to summary file
  • Optimized modified base aggregation

Bubbles

17 Nov 23:15
Compare
Choose a tag to compare

This point release includes some minor bug fixes and adds minor features and optimizations.

New features:

  • Remove requirement for per_read_mods output when requesting mod_mappings
  • Add script to output all motif hits to a BED file
  • Optimize per-site threshold command (for training modified base models)
  • Optimize standalone aggregation script when passing valid read IDs

Bug Fixes:

  • Fix bug in per-site threshold signal_mapping output

Crush

09 Nov 20:21
Compare
Choose a tag to compare

This point release includes some minor bug fixes and adds minor features.

New features:

  • New bedmethyl features (megalodon_extras merge aggregated_modified_bases)
  • Expit-scaled modified base aggregation (potentially useful for 5hmC Rerio model; #47)
  • More efficient per-site modified base annotation scripts.

Bug Fixes:

  • Fix per-read text output commands (fixes #68)

Marlin

20 Oct 18:46
Compare
Choose a tag to compare

This point release includes some minor bug fixes and adds minor features.

New features:

  • Sequence variant calibration files for Rerio modified base models
  • Improved documentation
  • Add API example script (in tests)
  • Optimize megalodon_extras calibrate generate_variant_stats command

Bug fixes:

  • Fix signal coordinates in mapping summary file
  • Fix alphabet printing bug

Gill

30 Sep 22:07
Compare
Choose a tag to compare

This a point release including a number of bug fixes and a few minor new features.

New features:

  • Include raw signal endpoints in reference mapping summary file
  • Add check for compatible Guppy and pyguppy versions
  • Minor improvements to megalodon_extra validate results command

Bug fixes:

  • Incorrect handling of mod_mappings output with signal_mapping options (Fixes #58)
  • Fix error when no modified base scores are added to database
  • Fix multiprocessing queues to correctly indicate full output queues
  • Fix potential bottleneck for modified base data storage

Dori

18 Sep 16:16
Compare
Choose a tag to compare

This release adds per-site modified base ground truth aided basecaller training data set creation. See docs.

This release also fixes a number of bugs:

  • Correct base quality scores in mod_mappings.bam output ( #56 )
  • --num-reads option functionality restored
  • Fixed bug where reads were skipped
  • Add threading and multiprocessing to read enumeration and raw signal extraction to recover more robust performance ( #49 )

Chum

08 Sep 19:12
Compare
Choose a tag to compare

This release includes two internal features to produce more robust processing speeds.

  1. Read batching

    • Each worker process sends a batch of reads (default 50) to the Guppy basecall server while processing results as they are returned to the client.
    • This should enable the Guppy basecall server to process reads more efficiently
  2. Signal extraction threading

    • Extract read IDs and raw signal using multiple threads (default 8).
    • This makes signal extraction more efficient on slower disk or when analyzing single read FAST5 format data.