Added new datasets sibship
and halfsib
.
In the app, these are available as Example1
and
Example2
, respectively.
The default genetic map, previously built-in to KLINK, has been
moved to the norSTR package under the name
map50
, and is now imported from there.
KLINK now employs the special lumping methods recently
implemented in pedprobr and pedmut.
This permits the use of complex mutation models
(e.g. Stepwise
) in most common pedigrees, including
paternity and full/half siblings. In larger pedigrees, nonlumpable
models are replaced with the simpler ‘Equal’ model.
The LR table shown in the app has a new look, now using the same layout and marker order as the other tables. The “LR table” sheet of the Excel download also uses this marker order. Note: The Report sheets has not changed.
New app controls Empty markers
(hide/show) and
Likelihoods
(hide/show/loglik) for the LR table in the app
and the Excel download.
New app control Decimals
for setting the number of
decimals in displayed tables.
Inconsistent marker names are handled more gracefully. For
instance, Penta D
, Penta.D
,
PentaD
, PENTA D
are now recognised as the same
marker (similarly for other markers, notably Penta E
and
vWA
).
The Marker data
table now excludes unused mutation
model columns (app only).
The Excel download includes a new sheet, Plots
,
containing the two pedigree plots.
Added tooltips explaining app controls and table columns.
Busy-indicator: A pulsating effect is now shown when the app is busy.
Renamed columns in the linkage map: Chrom
->
Chr
and PosCM
-> cM
. When
loading custom maps, variations of these names are accepted.
Only on shinyapp.io: Show banner warning about uploading sensitive data.
Empty the .xml
field when a new .fam
file is loaded.
RStudio now stops the app when the user closes the browser.
Graceful handling of input without data.
The Marker map
control has been moved up to the
Input
app region.
Fixed a bug appearing when the input file contains an unmapped single marker.
Fixed a bug giving slightly incorrect LR values in some cases
when a stepwise mutation model was used. This bug was related to
lumping, which is now handled differently (see entry under
New features
).
Modified the selection of markers for the
Unlinked report
. Within each pair, the marker with highest
PIC value is chosen. For consistency, the PIC values used in this
calculation are now always based on the same database
(NorskDB_2024).
Minor tweaks of app appearance.
This version is a major update of both the KLINK package and the shiny app.
Expanded built-in genetic map covering 50 common STR markers, up from 18. Unlike the previous version, where linkage pairs were hard coded, the markers are now dynamically paired up after data is loaded, ensuring more optimal use of the data.
The user may set a maximum linkage distance (cM), with the effect that markers farther than this are considered unlinked.
Improved karyogram plot for clearer visualisation of markers and linkage.
Improved readability of tables, using colours to indicate linked markers.
New button “Reset” resetting all fields of the shiny app.
New XML input field allowing the user to integrate data from other sources (particularly: Amelogenin genotypes and ID initials).
Update pedprobr version to 0.9.3, giving much faster likelihood calculations in many cases.
Show wait cursor during LR calculations.
Use scientific format also for small LR totals in-app.
Unlinked report: Prioritise D5S2500 over D5S2800 if both are present, and similarly D10S2325 over D10S1435, and D19S253 over D19S433 (reverses previous behaviour).
Fix typos in built-in linkage map.
Fix outdated mutation model in dataset
paternity
.
Internal reorganisation, moving the Shiny code to
inst/shiny/app.R
.
launchApp()
is now the main launching function, with
runKLINK()
as an alias.
Fixes a minor regression error in the previous version.
New button “Mask” for downloading a masked version of the dataset.
Reading/writing .fam
files is now handled by the new
package pedFamilias()
, which has been split out from
forrel
.
Added R option KLINK.debug
which can be set to TRUE
for debugging (only when running KLINK from R).
Use (experimental) autoScale
option in pedigree
plots.
Minor internal tweaks and bug fixes.
Although KLINK is primarily a Shiny app, the package now documents and exports the main functions, enabling analysis in R as well.
Added new dataset paternity
, with simulated data for
a paternity case (including a mutation).
Improved formatting in output Excel document.
Simplified code in sync with recent pedsuite updates. The plots may appear slightly modified.
Add karyogram showing marker positions.
Add sheets “Unlinked” and “Linked only” in excel output.
Tweaked button placements.
New button letting the user choose fallback mutation model (applied when a model specified in the input file fails for whatever reason).
Fixed a couple of minor bugs
Added a NEWS.md
file to track changes to the
package.
pedtools version 2.2.0 is now required.
If the pedigrees prohibit arbitrary lumping, all complex mutation models are replaced with a simpler model. A notification to this effect is added when loading the file. This behaviour is currently triggered if either pedigree has an untyped nonfounder.