NEWS.md

In oyvble/MPSproto: A Quantitative Model for MPS data with complex stutters

TODO:

• Optional pre-processing step for MPS sequence data: BayesHammer for sequence correction of noise.
• Optional to have a dynamic AT (per-sample)?

Suggestion: make a separate function called "getNoiseData"

MPSproto v0.9.7 (Release date: 2024-03-14)

• Fixed bug in GUI causing default minimum frequency to be zero (it should be the minimum observed instead). Also the minimum frequency based on database size did not apply.
• Updated plotMPS to make it possible to view reference profiles correctly when CE is given to MPS data (CE:STRING)

MPSproto v0.9.6 (Release date: 2023-11-24)

• Fixed GUI issue when earlier projects with calculations were loaded (adj.LL element not found)

MPSproto v0.9.5 (Release date: 2023-11-21)

• Improved implementation of calculating logLik per marker.
  • Obtained directly using new argument in calcLogLikC_prediction.R: First argument in loglikPrediction_allcomb.cpp is now a vector.

MPSproto v0.9.4 (Release date: 2023-11-03)

• Fixed AIC formula issue in GUI (gui.R). Renamed to "adj.logLik"

MPSproto v0.9.3 (Release date: 2023-08-23)

• Fixed bugs:

  • The degradation model was turned on even when the user had selected "No" when having kit selected (gui updated).
  • The program chashed for calculation under Hd when a reference had missing markers (getUpperLR updated).

• Updated to handle missing evidence markers (prepareData_prediction,prepareC_prediction).

MPSproto v0.9.2 (Release date: 2023-08-23)

• Avoid program to crash when using degradation model and having kit selected, and "CE:" prefix format is not used.
• Improved robusness of generation of noise in genMPSevidence.
• An error is given if the number of alleles and reads are not the same.

MPSproto v0.9.1 (Release date: 2023-05-11)

• Fixed bug when assigning values in toolbar in GUI. -> Adding "ncolumns=1" at gui.R:L279.

MPSproto v0.9.0 (Release date: 2022-08-23)

• Including a GUI for interpretation (gui.R)

MPSproto v0.8.1 (Release date: 2022-07-29)

• Fixed bug in getStutterData causing stutter from homozygous genotypes to be removed.

MPSproto v0.8.0 (Release date: 2022-03-15)

• Implementing Negative Binomial (NB) as an model option in inferEvidence function: model=("GA,"NB")
  • Following functions changed: calcLogLikC_prediction, prepareC_prediction, prefit_prediction, validMLE, calcLogLikR_prediction
  • C++ now utilizing the function pnbinom_mu for cumulative calculation of NB
• genMPSevidence: Degradation not possible when generating evidence for MPS

• fitgammamodel renamed as fitSUMmodel which fits all implemented models (GA/NB).

• Updated functions for stutter calibration:

  • getStutterData function is more comprehensive (also take minStuttOccurence argument)
  • getFilteredData replaced by getStutterData (deprecated)
  • inferStuttRegModel renamed to inferStutterModel

MPSproto v0.7.1 (Release date: 2022-03-02)

• Removed function getComplement (not used)

MPSproto v0.7.0 (Release date: 2022-02-02)

• Included tutorial and scripts for calibrating data from paper

MPSproto v0.6 (Release date: 2022-02-14)

• Now also supporting CE format (Stutter model for BLMM can be included): getStutterIndex updated
• Take into account that getStutterIndex returns NULL if none of the alleles can have a certain stutter type
• Added functions:
  • getStutterTypeRule: Get motif change and motif rank for given stutter type
  • genMPSevidence: Used to generate MPS mixture based on a calibrated model

MPSproto v0.5 (Release date: 2022-01-12)

• Updated inferStuttRegModel function to also include validation plots. Also supports precision expression.

MPSproto v0.4 (Release date: 2021-12-03)

• n0 stutters are defined differently (getStutterIndex changed).
• normalising/minFreq argument can be modified by including it into the hypotheses object

MPSproto v0.3 (Release date: 2021-11-12)

• Including inferEvidence2 function to also estimate marker efficiency, penalized with it's defined prior.
• Increase the theoretical maximum for validMLE (take into account marker efficiency)
• Including helpfunctions for easy extraction of Param, LR, AIC.
• Fixed bug in plotMPS: Marker specific AT/ST argument was not shown.

MPSproto v0.2 (Release date: 2021-11-03)

• Changing NoiseSize model to be discrete instead of continuous (still Pareto).
• inferMarkerEfficiency function can now directly perform MCMC if mleObj argument is provided.
• Including logliki function to obtain per-marker logLikelihood.
• Including plotTopMPS for checking model (also gives stutter contribution per contributors)
• Including getStutterExpectations to obtain expected stutter proportions between each alleles in evidence data

MPSproto v0.1 (Release date: 2021-10-19)

• init release

oyvble/MPSproto documentation built on March 19, 2024, 5:32 a.m.