exSTRa v0.90.0

This release adds these features:

• Added known repeat databases for hg38 and GRCh38
• Known repeat database: name of start and end column more flexible now, so now the column name should end with either 'start' or 'end' (case-insensitive, though only the latest version of the Perl module is also case-insensitive). This means that any future known repeat files do not confusingly have the column names hg19_start and hg19_end when on a different genome build.

Under the hood changes:

• deleting old unused files and code
• testing tsum_test reproducibility (for any future optimisations)

exSTRa v0.89.0

Update repeat expansion disorders database

• Create separate files for hg19 and GRCh37
• Add several new loci

exSTRa v0.88.6

ggplot2 methods work in a limited capacity for exstra_score objects, incorporating x$data and x$sample table columns.

More bug fixes

• plotting the resulting tsum_test() test object no longer tries to access a global tsum variable.
• tsum_p_value_summary() correction parameter works
• some manual page corrections

Some work on making parallel tsum_test() work better, but with only minor benefit.

exSTRa v0.88.4

• New citation used in README and vignette
• Added some warnings
  • If data.table package is loaded after exSTRa
  • If parallel option is used in tsum_test(), as this is not currently beneficial due to overhead (after last years serial optimizations)
• Code cleanup to remove some build warnings

exSTRa v0.88.3

Updated citation to American Journal of Human Genetics. Additionally, some code cleaning and removing parallel package from vignette example.

exSTRa v0.88.2

A minor speedup (~1.2x) to tsum_test(). This gives a total speedup of approximately 200x from v0.87.

exSTRa v0.88.0

The new and improved exSTRa, with significant optimizations and case-control functionality

• Added link to brain expressed database files
• Added example script to generate exSTRa repeat genome wide
• A significant speedup to tsum_test(), at least 120x faster
• Further speedup to tsum_test() by implementation of early termination of simulation procedure when the p-value standard deviation is much smaller than the smallest p-value itself (total 170x speedup, or possibly even more with more simulations (B) or few true positive samples)
• Case-control testing in tsum_test() is now implemented
• Access to the old tsum_test for comparative purposes is accessible by exSTRa:::tsum_test_1

Note that these optimizations now mean that benefits of using tsum_test(..., parallel = TRUE) are somewhat reduced. This may be explored in the future to make parallel computation more beneficial.

exSTRa v0.87

Small update of documentation

• BioRXiv citation updated to revised preprint posted 25th July 2018
• Readme and example update, directing users to the vignette for an example

exSTRa v0.86

• P-values from simulation now include standard deviation (p.value.sd). When x is the output from tsum_test(), access by either x$stats or p_values(x).
• in tsum_test() and p_values(), correction by the number of loci or samples is now clearer ("locus" removed as an option - this was confusing)
• Simulation results during tsum_test() are reused for a slight speedup (~5%). Code requires a rewrite for a significant speedup here. The default number of simulations is now a factor of 10 less (9999 to 999), and script with B set will be slower without also changing B.

Bug fixes:

• groups.samples now works correctly when reading in STR data from Bio::STR::exSTRa (thanks to a113n)

exSTRa v0.84

Documentation update. Several more manual pages are written now, with some existing manual pages updated too.