v0.3.1rc1

Fixes

• [find-motifs] Bug where error would be reported when output tables are specified. Fixes #195

Adds

• [extract] Allows streaming read calls table directly to stdout.

v0.3.0

Fixes

• [validate] Fix bug where observations were not being balanced correctly.
• [dmr] Revert noodles to 0.50.0, partially fixes #178.

Adds

• [find-motifs] Add command to find motif sequences that are enriched for base modification.
• [entropy] Add command to calculate methylation entropy.
• [dmr] Add --segment option to categorize modified positions into contiguous "same" and "different" groups.
• [pileup] Add --header option to emit a column header.

Deprecations

• [pileup] Change output to always be all-tab delimited, next version will error with --only-tabs flag.

v0.2.8-rc1

Fixes

• [dmr] Don't crash when using matched samples, but not all samples have modification counts at a position. Fixes #165.
• [update-tags] Correctly handle implicit tags when there are no probs, i.e. C+h -> C+h.., fixes #160.

v0.2.7

Fixes

• [dmr] Header was incorrect with multiple samples
• [pileup] Improve performance when using --include-bed, only process contigs in the BED file.
• [dmr, single-site] When using multiple samples, don't fail a position when one or more samples doesn't have a modification call at that position.
• [extract] Expose queue size to reduce memory usage with long reads.
• [validate] Report number of calls filtered out with thresholds.

v0.2.6

Fixes

• [dmr, single-site] Don't require that there are equal numbers of samples for single site DMR with multiple samples. Fixes #140.
• [dmr, pairwise, region] Protect when zero bedmethyl records are found for a region, fixes #146.

Adds

• [validate] Adds on-the-fly filtering of reads by alignment identity and/or alignment length.

v0.2.5

Fixes

• [extract] Only emit mapped reads when --region is provided, but still emit unmapped bases in those reads unless --mapped-only is passed.
• [extract] Performance improvement due to better tracking of interval boundaries.
• [repair] Updates the MN tag on repaired records.

Adds

• [dmr, single-site] Refactor dmr pair without regions (i.e. single site analysis) to increase performance.
• [dmr, single-site] Add estimated MAP-based p-value to output.
• [all] Allows BED3 input for all options that use --include-bed. Strand will be assumed to be BOTH (equivalent to '.').
• [extract] Increases the kmer size limit to 50.

v0.2.5-rc2

Fixes

• [all] Reads with entirely implicit canonical calls are no longer skipped for "modbase info empty" or similar.

Adds

• adds the --no-implicit-calls flag to modkit update-tags so when changing from the "probability of modification" modes (".", and "") to explicit mode (?) implicit cannonical calls can be skipped. This is important, for example, if the modification calling model did not have the ? mode available.

v0.2.5-rc1

Adds

• modkit validate sub-command for rigorous testing of modified base calling models when ground truth labels are known.

Fixes

• [all] Improve performance when using commands on transcriptome reference (or any reference with many sequences less than ~100kb).

v0.2.4

Adds

• [extract, adjust-mods, update-tags, call-mods] Parse MN tag in order to use secondary and supplementary alignments.

Fixes

• [all] Improve performance slightly when using short and frequent motifs with --motif option.

v0.2.3

Adds

• [dmr, multi] Allow site-level scoring by omitting the --regions argument. Sites will be collected from the input bedMethyl files.
• [dmr] Friendlier handling of missing files and when regions aren't found in the bedMethyl input files.
• [dmr] allow filtering on valid coverage without changing input
• [extract] output "read calls" extract table, a TSV with the base modification calls for every modified position in every read using the same thresholding algorithm as pileup
• [extract] allow filtering of calls by reference motif (like in pileup) as well as BED regions (and exclude regions)

Fixes

• [extract] Improve performance, especially on longer reads.
• [extract] Improve performance with long reads (actually a bug fix)
• [extract] num_soft_clipped_start and num_soft_clipped_end were incorrect on some reverse-mapped reads