4.1.0

Fixes bugs and tweaks outputs, including:

• Removes --analysis-mode fast from pangolin call so it uses usher (by default)
• Adds sequencing_tech and sample_sequencing_datetime fields to qc_and_lineages summary file * (and downstream reports)
• Changes "stringent" condition to use "overall_fail" (more exclusive) rather than "any_fail"
• Changes logging to include stdout instead of just stderr
• Adds retry for AWS downloads because of intermittent failures
• Fixes search-and-replace error where "qc" was replaced by "src"

4.0.1

• Fix typo preventing aggregation of pi metrics at QC step
• Add more detail about cluster creation to README.md

4.0.0

Major release:

• Change from SGE scheduler to Slurm scheduler to support AWS ParallelCluster version 3
• Update installation script and conda environment yaml, including
  • Upgrade to pi_from_pileup version 1.0.3
  • Install iVar via conda
• Update and extend README.md with detail on installation and cluster creation
• Update folder structure and script names for better usability
• Output full *pi_from_pileup.tsv for each sample in addition to *pi-metric.tsv

3.9.0

• Sets Ready for release? value to No in bjorn-ready file for any sample with known bad metadata
• Generates md5 checksums for each run's *-passQC.fas and *-summary.csv file during QC step
• Creates a file of the md5 checksums of each *-passQC.fas and *-summary.csv file used as input during the lineage identification step
  • Copies of each *-passQC.fas and *-summary.csv are no longer uploaded to the S3 output, to reduce duplication and storage use
• Removes creation of no-longer-used passQC output folder during lineage identification
• Refactors confusing code in custom stats generation QC step

3.8.0

• Adds ability to input bams produced by Genexus instrument instead fastqs
• Adds variants_qc functionality that runs both per-sample processing and QC summarization for single sequencing run
• Corrects bug in which Pct >=Q30 value for single-end reads was output as fraction instead of percent
• Corrects bug in which uncapped_reads_lt_100k metric was actually calculated as uncapped_reads < 50k
• Modifies QC metric calculations to account for metrics that are missing because they cannot be calculated from bam inputs.
  • A missing metric is no longer treated as a False value for that metric; metrics are now allowed to be True, False, or None
  • None values are ignored when calculating logical combinations such as any_fail (except in the case where ALL metrics are None, in which case the logical combination of them is treated as False).

3.7.0

• Updates sample names in unit tests to follow naming convention
• Moves regression test files out of this project
• Moves helper scripts out of this project

3.6.0

Support for pangolin 4+ :

• Adds --analysis-mode fast to pangolin call to continue to use pangoLEARN as classifier
• Removes consensus sequence munging no longer used in pangolin 4+
• Updates expected of status column name in pangolin output

As of this release, C-VIEW is not compatible with pre-4.0 pangolin versions. This should not be a problem as C-VIEW automatically updates pangolin to the latest on every run, so there should be no case in which it is attempting to run an older version.

3.5.0

• Removes tree-building and visualization functionality (now performed externally)
• Includes stand-alone update_files_w_iqtree_renames.py script to update alignment and metadata files with IQ-TREE-munged sequence names (to support insertion of new sequences into an IQ-TREE-created base tree)
• Documents cumulative_pipeline function
• Adds -J option to samtools depth to exclude deletions from coverage calculations
• Excludes samples with negative ages from bjorn file

3.4.1

• Updates git repo link and conda environment names in install.sh
• Explicitly deactivates pangolin environment and reactivates main C-VIEW environment after pangolin use in lineages.sh
• Removes code to support "loose" samples subset processing in lineages.sh

3.4.0

Modifies creation of bjorn summary to:

• add a column containing the bam s3 url
• prevent an error if no samples are eligible for bjorn
• mark a sample as ready for release if it has overall_fail = False and has new metadata column submit_to_gisaid = True (this should prevent e.g. CDPH sample inclusion)
• add the state abbreviation into the virus name

Modifies tree prep to:

• define "stringent" samples as those with any_fail = False (instead of usable_for = variant_and_epidemiology)
• cease creating metadata and fas files for "loose" samples (such are no longer defined)