Changelog • archR

New features

Custom ordering of individual clusters within collated clusters. Earlier this followed label ordering as per cutree output
Improved messages for model selection using cross-validation
(User-facing) The agglomeration and distance method used for collation of clusters can be set as part of configuration settings

Breaking changes

(User-facing) Argument name inner_chunk_size changed to chunk_size in archR_set_config()

New features

Accounting and correcting for over-fitting improves clusters/architectures’ quality
Improved messages
(User-facing) New function collate_clusters()

Breaking changes

(User-facing) tolerance no longer an argument in archR_set_configuration

New features

Can choose to fix aspect ratio for visualization functions
Ability to detect occurrence of just for sake clustering when fetching from HAC

New features

(User-facing) Getter functions to fetch archR result object elements.
- seqs_str() to fetch DNA sequences for a particular cluster, iteration etc.
- get_clBasVec() to fetch basis vectors element at given iteration.
- get_clBasVec_k() to fetch number of basis vectors/clusters.
- get_clBasVec_m() to fetch basis vectors.
- get_seqClLab() to fetch cluster Ids for each sequence.

Breaking changes

(User-facing) Arguments of all functions consistently use underscores.

Bug fixes

Fix staged installation error. This enables seamless installation on macOS.

New features

CI with GHA.
(User-facing) archR result list has new elements:
- timeInfo if timeFlag is TRUE;
- clustSol storing combined clusters from the last iteration of archR.
(User-facing) Architectures/clusters sequence logos can be plotted with auto y-axis limits for information content (0-max instead of 0-2).
(User-facing) TFBSTools moved from dependency to suggests. This eases installation of archR by reducing possibility of failed installation.

New features

Accepts FASTA sequences as DNAStringSet object(s)