ASEFilter-class: SpliceWiz filters to remove low-confidence alternative...

In alexchwong/SpliceWiz: interactive analysis and visualization of alternative splicing in R

SpliceWiz filters to remove low-confidence alternative splicing and intron retention events

Description

SpliceWiz implements a number of novel filters designed to exclude alternative splicing events (ASEs) that yield low-confidence estimates.

Usage

ASEFilter(
  filterClass = c("Data", "Annotation"),
  filterType = c("Depth", "Participation", "Consistency", "Modality", "Protein_Coding",
    "NMD", "TSL", "Terminus", "ExclusiveMXE", "StrictAltSS"),
  pcTRUE = 100,
  minimum = 20,
  maximum = 1,
  minDepth = 5,
  condition = "",
  minCond = -1,
  EventTypes = c("IR", "MXE", "SE", "A3SS", "A5SS", "AFE", "ALE", "RI")
)

Arguments

filterClass	Must be either "Data" or "Annotation". See details
filterType	Must be a valid "Data" or "Annotation" filter. See details
pcTRUE	If conditions are set, what percentage of all samples in each of the condition must satisfy the filter for the event to pass the filter check. Must be between 0 and 100 (default 100)
minimum	Filter-dependent argument. See details
maximum	Filter-dependent argument. See details
minDepth	Filter-dependent argument. See details
condition	(default "") If set, must match the name of an experimental condition in the NxtSE object to be filtered, i.e. a column name in colData(se). Leave blank to disable filtering by condition
minCond	(default -1) If condition is set, how many minimum number of conditions must pass the filter criteria. For example, if condition = "Batch", and batches are "A", "B", or "C", setting minCond = 2 with pcTRUE = 100 means that all samples belonging to two of the three types of Batch must pass the filter criteria. Setting -1 means all elements of condition must pass criteria. Set to -1 when the number of elements in the experimental condition is unknown. Ignored if condition is left blank.
EventTypes	What types of events are considered for filtering. Must be one or more of c("IR", "MXE", "SE", "A3SS", "A5SS", "AFE", "ALE", "RI"). Events not specified in EventTypes are not filtered (i.e. they will pass the filter without checks)

Details

Annotation Filters

• Modality: Filters for specific modalities of ASEs. All events belonging to the specified EventTypes are removed. No additional parameters required.

• Protein_Coding: Filters for alternative splicing or IR events involving protein-coding transcripts. No additional parameters required.

• NMD: Filters for events in which one isoform is a predicted NMD substrate.

• TSL: filters for events in which both isoforms have a TSL level below or equal to minimum

• Terminus: In alternate first exons, the splice junction must not be shared with another transcript for which it is not its first intron. For alternative last exons, the splice junction must not be shared with another transcript for which it is not its last intron

• ExclusiveMXE: For MXE events, the two alternate casette exons must not overlap in their genomic regions

• StrictAltSS: For A5SS / A3SS events, the two alternate splice sites must not be interupted by detected introns

Data Filters

• Depth: Filters IR or alternative splicing events of transcripts that are "expressed" with adequate Depth as calculated by the sum of all splicing and IR reads spanning the event. Events with Depth below minimum are filtered out

• Participation: Participation means different things to IR and alternative splicing.
  
  For IR, Participation refers to the percentage of the measured intron covered with reads. Only introns of samples with a depth of intron coverage (intron depth) above minDepth are assessed, where introns with coverage percentage below minimum are filtered out.
  
  For non-IR ASEs, Participation refers to the percentage of all splicing events observed across the genomic region (SpliceOver metric) that is compatible with either the included or excluded event. This prevents SpliceWiz from doing differential analysis between two minor isoforms. Instead of IntronDepth, in AS events SpliceWiz considers events where the SpliceOver metric exceed minDepth. Then, events with a SpliceOver metric below minimum are excluded.
  
  We recommend testing IR events for > 70% coverage and AS events for > 40% coverage as given in the default filters which can be accessed using getDefaultFilters
  
  

• Consistency: Skipped exons (SE) and mutually exclusive exons (MXE) comprise reads aligned to two contiguous splice junctions. Most algorithms take the average counts from both junctions. This will inadvertently include transcripts that share one but not both splice events. To check that this is not happening, we require both splice junctions to have comparable counts. This filter checks whether reads from each splice junction comprises a reasonable proportion of the sum of these reads.
  
  Events are excluded if either of the upstream or downstream event is lower than total splicing events by a log-2 magnitude above maximum. For example, if maximum = 2, we require both upstream and downstream events to represent at least 1/(2^2) = 1/4 of the sum of upstream and downstream event. If maximum = 3, then each junction must be at least 1/8 of total, etc. This is considered for each isoform of each event, and is NOT tested when total (upstream+downstream) counts belonging to each isoform is below minDepth.
  
  IR-events are also checked. For IR events, the upstream and downstream exon-intron spanning reads must comprise a reasonable proportion of total exon-intron spanning reads.

We highly recommend using the default filters, which can be acquired using getDefaultFilters

Value

An ASEFilter object with the specified parameters

Functions

• ASEFilter(): Constructs a ASEFilter object

See Also

Run_SpliceWiz_Filters

Examples

# Create a ASEFilter that filters for protein-coding ASE
f1 <- ASEFilter(filterClass = "Annotation", filterType = "Protein_Coding")

# Create a ASEFilter that filters for Depth >= 20 in IR events
f2 <- ASEFilter(
    filterClass = "Data", filterType = "Depth",
    minimum = 20, EventTypes = c("IR", "RI")
)

# Create a ASEFilter that filters for Participation > 60% in splice events
# that must be satisfied in at least 2 categories of condition "Genotype"
f3 <- ASEFilter(
    filterClass = "Data", filterType = "Participation",
    minimum = 60, EventTypes = c("MXE", "SE", "AFE", "ALE", "A3SS", "A5SS"),
    condition = "Genotype", minCond = 2
)

# Create a ASEFilter that filters for Depth > 10 in all events
# that must be satisfied in at least 50% of each gender
f4 <- ASEFilter(
    filterClass = "Data", filterType = "Depth",
    minimum = 10, condition = "gender", pcTRUE = 50
)

# Get a description of what these filters do:
f1
f2
f3
f4

alexchwong/SpliceWiz documentation built on March 17, 2024, 3:16 a.m.