NEWS

---

INSPEcT 1.16.1

• added the slot 'version' that store the version used in each object created (IMPORTANT! objects without this slot, i.e. created before this version cannot be used anymore and must be regenerated)
• added 2 wrap-up function to run the entire INSPEcT pipeline with a single command line
• ratePvals modified (calculated at the time of modeling via calculateRatePvals method and stored into slot ratePvals of the class INSPEcT)
• changes in model params accession (modeling parameters are set via modelRates and only accessed via modelingParams; model selection parameters are set via calculatePvals and only accessed via modeingParams)
• modifications to INSPEcT-GUI (Added Fix-Y-Axis & pvals in Y-labels, Confidence intervals controlled by a button rather than the checkbox)
• converted gene class names from a,c,b for synthesis, processing and degradation to s,p,d

INSPEcT 1.16.0

• added processingDelay methods and 2 shiny apps (ProcessingRateDelay and INSPEcT-GUI)
• added INSPEcT-GUI vignette
• New +Nascent modeling based on confidence intervals, new -Nascent derivative modeling, no more time transformation, new simulated data genesis, new data corruption, new HIB model selection, new graphical functions (rocCurve comparative and correlationPlot).
• added non-functional modeling (modelRatesNF) and the method makeOscillatorySimModel
• simulated datasets are now generated with contamination from total to nascent RNA fraction (default contamination values: b=0.3 and noise parameter=4)
• modifications on compareSteadyNoNascent method (arguments trivialAngle, returnScores and referenceCondition added in order to define the trivial angle without extracting from the dataset, to return the scores instead of the classification and to define a reference condition instead of using the median as the reference)
• added the comparative mode in the rocCurve method and the p=0.05 point

INSPEcT 1.14.0

• added labeled + pre-existing RNA mode when dedDuringPulse is OFF
• limitation of complexity added

INSPEcT 1.12.1

• major update of INSPEcT, which now handles additional functions and analysis, such as: (i) extracting RNA dynamics from only total RNA-seq data, (ii) assessing differential regulation between steady states with an improved statistics (compareSteady method),
• minor updates regard the estimation of the variance from replicates, that now uses the package "DESeq2" (when read counts are available) or "plgem" (when only expression data is available) ######## the INSPEcT datasets older than this version are NOT COMPATIBLE anymore ############## ######## and must be updated using the function "convert_ids", available in the ############## ######## "inst" folder of the package. Usage: ############## > source(file.path(system.file(package='INSPEcT'), 'convert_ids.R')) > new_ids <- convert_ids(old_ids, degDuringPulse=c(FALSE,TRUE))

INSPEcT 1.5.5

• updated the results on the simulated data in the vignette

INSPEcT 1.5.4

• modified the argument strandSpecific in makeRPKMs function, so that now the user can perform strand-specific read counting with this possible modes: 0 => unstranded 1 => stranded 2 => reversely stranded

INSPEcT 1.5.3

• modified internal functions inferKBetaFromIntegral, inferKBetaFromIntegralWithPre, inferKGammaFromIntegral in order to reduce the number of the missing values (NA) in the output

INSPEcT 1.5.2

• Updated the documentation

INSPEcT 1.5.1

• Fixed a bug that caused a mis-choiche of sigmoid or impulse function during modeling

INSPEcT 1.1.4

• Added capabilities for the comparison of two steady-state conditions

INSPEcT 1.1.3

• INSPEcT is now independent from pyhton and HTSeq for counting reads in intronic and exonic regions and now uses the GenomicAlignments and Rsamtools Bioconductor packages functionalities. Thanks to this modification INSPEcT is now fully contained into the R/Bioconductor framework, being therefore more complant with Bioconductor guidelines. Additionally BAM files can be directly used, without the time consuming and memory consuming step of BAM->SAM conversion. Methods related to the creation of the annotation of intronic and exonic features in GTF format to provide to pyhton are now dismissed consequently and will be probably introduced as methods in the Bioconductor package 'compEpiTools'

INSPEcT 1.1.2

• Parallel computation is now managed completely within the BiocParallel framework. "newINSPEcT" function and "modelRates" method take as input the argument BPPARAM which is used to handle the parallelization. By default, BPPARAM is assigned with bpparam() function of BiocParallel package, which guarantee the maximum number of available cores used and the usage of forking in Linux and MacOS-X and the usage of the package Snow for Windows machines.
• nCores methods and arguments are now deprecated.

INSPEcT 1.1.1

• Re-introduced inferKBetaFromIntegralWithPre, which disappeared in the devel version following 1.0.1 (excluded)
• selection of best model is now done applying brown test on the pairs of model where at least one of the two has a chi-squared test lower than the threshold. This is done because in case only one rate leads the dynamics, all the model which don't involve that rate won't have low chi-squared and no comparison will be made. This leads to brown p-values of 1 on that specific rates (change in method "ratePvals")
• in newINSPEcT, the guess of new rates can be done without assuming that degradation does not occur during the pulse
• Solved two problems. One that occurred during modeling for genes with estimated variance within replicates equal to zero: in these cases the variance is estimated within the time course. A second problem was encountered in the parameter initialization of impulse model: h1 cannot be zero in order to evaluate a finite value."
• Evaluate 'modelRates' within the vignette in parallel only in Linux and Dawin environments. This is done to avoid timeout in the build process on Bioc servers.
• Better estimation of rates in case degDuringPulse=TRUE (newINSPEcT function). Added controls on input arguments in newINSPEcT function. Fixed a bug in the saturation of values out of breaks in inHeatmap method (this change could cause a different clustering order of genes in the heatmap). Added the palette argument to inHeatmap method.
• Fixed a bug in '[' method and unpdated the NAMESAPACE and DESCRIPTION files according to the update of the 'unlist' method that is exported from BiocGenerics and not from GenomicRanges anymore.