ChIP-seq candidate-peak significance (MACS dynamic lambda)

For a candidate peak of the given width with k ChIP tags, the uniform background rate lambda_BG is replaced by the dynamic \(\lambda_{local} = \max(\lambda_{BG}, [\lambda_{1k},] \lambda_{5k}, \lambda_{10k})\), where \(\lambda_{Xk}\) is the control tag count of the X-kb window centred on the peak rescaled to the peak width (count * width / X000). When no control sample exists the 1 kb term is omitted (use_1k = FALSE). The peak p-value is the Poisson upper tail \(P(X \ge k)\) under lambda_local and fold_enrichment is k / lambda_local.

Usage

Chipsq(
  count,
  width,
  lambda_bg,
  count_1k = NULL,
  count_5k = NULL,
  count_10k = NULL,
  use_1k = TRUE
)

Arguments

count

ChIP tag count per candidate peak.

width

Peak region width in bp.

lambda_bg

Expected background tag count for the region.

count_1k, count_5k, count_10k

Control tag counts in the centred 1 kb, 5 kb and 10 kb windows (optional).

use_1k

Set FALSE when no control sample exists.

Value

List with pvalue, lambda_local, fold_enrichment, count, width, n_peaks, method.

References

Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nusbaum, C., Myers, R. M., Brown, M., Li, W. and Liu, X. S. (2008), Model-based Analysis of ChIP-Seq (MACS), Genome Biology 9(9), R137. Dynamic lambda and Poisson p-value, Methods, Peak detection.

Examples

Chipsq(count = c(1, 2, 3, 4, 5, 6, 7, 8), width = 5L, lambda_bg = 5L)
#> $pvalue
#> [1] 0.9932621 0.9595723 0.8753480 0.7349741 0.5595067 0.3840393 0.2378165
#> [8] 0.1333717
#> 
#> $lambda_local
#> [1] 5 5 5 5 5 5 5 5
#> 
#> $fold_enrichment
#> [1] 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
#> 
#> $count
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $width
#> [1] 5 5 5 5 5 5 5 5
#> 
#> $n_peaks
#> [1] 8
#> 
#> $method
#> [1] "MACS dynamic-lambda Poisson peak test (Zhang et al. 2008)"
#>