\title{Class \code{"BamTallyParam"}}
  A \code{BamTallyParam} object stores parameters for
  \code{\link{bam_tally}}. The function of the same name serves as its

  BamTallyParam(genome, which = GRanges(),
                desired_read_group = NULL,
                minimum_mapq = 0L,
                concordant_only = FALSE, unique_only = FALSE,
                primary_only = FALSE, ignore_duplicates = FALSE,
                min_depth = 0L, variant_strand = 0L, variant_pct = 0,
                ignore_query_Ns = FALSE,
                indels = FALSE, min_softclip = 0L, max_softclip = 0L,
                exon_iit = NULL, IIT_BPPARAM = NULL,
                xs = FALSE, read_pos = FALSE,
                min_base_quality = 0L, noncovered = FALSE, nm = FALSE)
  \item{genome}{A \code{GmapGenome} object, or something coercible to one.}
  \item{which}{A \code{IntegerRangesList} or something coercible to
    one that limits the tally to that range or set of ranges. By
    default, the entire genome is processed.
  \item{desired_read_group}{The name of the read group to which to limit
    the tallying; if not NULL, must be a single, non-NA string.}
  \item{minimum_mapq}{Minimum mapping quality for a read to be counted
    at all.}
  \item{concordant_only}{Consider only what gnsap
    calls \dQuote{concordant} alignments.
  \item{unique_only}{Consider only the uniquly mapped reads.}
  \item{primary_only}{Consider only primary pairs.}
  \item{ignore_duplicates}{Whether to ignore the reads flagged as
    PCR/optical duplicates.
  \item{min_depth}{The minimum number of reads overlapping a position for
    it to be counted.}
  \item{variant_strand}{The number of strands on which a variant must be
    seen for it to be counted. This means that a value of 0 will report
    reference alleles in addition to variants. A value of 1 will report
    only positions where a variant was seen on at least one strand, and
    2 requires the variant be seen on both strands. Setting this to 1
    is a good way to save resources.}
  \item{variant_pct}{The minimum alternate allele fraction
    for a variant to be reported for a strand.} 
  \item{ignore_query_Ns}{Whether to ignore the N base pairs when
    counting. Can save a lot of resources when processing low quality data.}
  \item{indels}{Whether to return indel counts. The \code{ref} and
    \code{alt} columns in the returned \code{VRanges} conform to VCF
    conventions; i.e., the first base upstream is included. The range
    always spans the sequence in \code{ref}; so e.g. a deletion extends
    one nt upstream of the actual deleted sequence.
  \item{min_softclip, max_softclip}{
    Minimum and maximum length of soft clips that are
    considered for counting. Soft-clipping is often useful (for GSNAP at
    least) during alignment, and it should be preserved in the
    output. However, soft clipping can preferentially occur in regions
    of discordance with the reference, and if those clipped regions are
    ignored during counting, the allele fraction is misestimated.
  \item{exon_iit}{An object which indicates the exons to be used for
    tallying codons (a character value indicating an existing .iit file, a
    \code{GRangesList} of exons by gene or a \code{TxDb} object from which
    to make such a \code{GRangesList}) or \code{NULL} indicating no
    codon-level tallying should be done.}
  \item{IIT_BPPARAM}{A \code{BiocParallelParam} object to use when
    generating the iit file from an R object. Ignored if \code{exon_iit}
    is a character vector or \code{NULL}
  \item{xs}{Whether to tabulate reads by XS tag, the aligner's best
    guess about the strand of transcription.
  \item{read_pos}{Whether to tabulate by read position.
  \item{min_base_quality}{Minimum base quality cutoff. Calls of lower
    quality are not counted, except in the total raw depth.
    Whether to report zero tallies, where there is no coverage.
    Whether to tally by NM tag, the number of mismatches for a read.