1.7.2

Fixing fix of 1.7.1

1.7.1

Bugfix

Fixing some issues with blast report generation when using newer BLAST databases. Source of the problemis still unclear but appears to be linked to new realeses of the BLAST db and concerns only specific taxids or sequences.

1.7.0

1.7.0

Breaking changes

Older configuration files are not compatible anymore. Updating the files by adding the following line
should yield the same results as before:

blast_filter_low_complexity: True

New features

It is now possible to deactivate the default low-complexity filter of the BLAST search.
This can be advantageous if you expect your barcode to contain low-complexity sequences which could
prevent getting any match at all.
This behaviour can be activated/deactivated by changing the blast_filter_low_complexity from True to False.

The default behaviour (True) uses the default 'DUST' filter of the blast tool:
-dust 20 64 1 -soft_masking true.
Switching the filter off runs blast with the modified parameters:
-dust no -soft_masking false

1.6.6

Fixes

• Corrects parsing of the trim_primers_3end parameter (#64)
• Added a primer disambiguation step that converts primers sequences in the IUPAC ambiguous nucleotide format to their explicit forms (#63)

1.6.5

1.6.5

Fixes

• BLAST rule now correctly uses the threads_sample parameter instead of threads. This results in better ressource management for the BLAST rule.

1.6.4

1.6.4

Benchmarking paper for 16S Metabarcoding of meat products is online:

Denay, G.; Preckel, L.; Petersen, H.; Pietsch, K.; Wöhlke, A.; Brünen-Nieweler, C.
Benchmarking and Validation of a Bioinformatics Workflow for Meat Species Identification Using 16S rDNA Metabarcoding.
Foods 2023, 12, 968. https://doi.org/10.3390/foods12050968

New features

• It is now possible to filter specific sequences form the database using a list of accession provided in a text file using the seq_blocklist parameter (#60).

Improvements

• Taxa names are now displayed in the benchmarking report
• Fixed many typos and errors in the documentation
• Improved report aesthetics
• Improved the fetch_nt_blast.sh script to make it easier to resume interupted processes, also more robust

Fixes

• Fixed confusion matrix calculation when the expected taxid rank is above the target rank

Foodme v1.6.3

1.6.3

Improvements

• Benchmarking: confusion table now reports prediction rank as well

Bug fix

• Fixed major problem in confusion matrix determnination for the benchmark module. Prior to this fix, False negatives were not correctly reported.
• Small fix to report aggregation rules for a rarely happening failure
• Fixed composition summary when input is empty

foodme v1.6.2

Bug fix

• Moved the benchmark module form a rule to a workflow. THis allows to ignore the benchmark arguments when running routine
  analysis, as originally intended. Benchmarking is now called with the -s path/to/FooDMe/woorkflow/benchmark argument.
  This has no impact on the basic analysis (no -s argument) or the paramspace analysis (-s path/to/FooDMe/woorkflow/paramspace).

Documentation

• Fixed a few errors in the config.yaml comments
• Modified documentation of the benchmark module

foodme v1.6.1

Bug fix

• Fixed handling of last common node calculation in confusion matrix where last common ancesotr is the root node.
• Modified dependencies in environement definition files to solve some issues in conda solving

foodme v1.6.0

This update is not backwards compatible.
A configuration file update is nescessary.

New features

• Added a new Snakefile for parameter space exploration.
  Basically acts as a wrapper around the foodme benchmark workflow
  for parameter grid search using snakemake's Paramspcae utility.

Non backward compatible changes

• Flattened the parameter structure in the configuration.
  This is more compatible with the --config CLI argument and
  was required for the implementation of the parameter space exploration workflow.
  This requires users to update their configurations.

Documentation

• Added documentation for the paramspace workflow.