In Bishop-Laboratory/RLSeq: RLSeq: An analysis package for R-loop mapping data

r i. r names(resAnchor) {#r resAnchor}

pred <- rlresult(object, resultName = "predictRes")
verd <- ifelse(
  pred$prediction == "POS", 
  "<strong style='color: #602D64'>\"POS\"</strong> (i.e., robust R-loop mapping)", 
  "<strong style='color: red'>\"NEG\"</strong> (i.e., poor R-loop mapping)"
)

Predicted label for sample r object@metadata$sampleName is r verd.

Details

Additional Details

To evaluate sample quality, a binary classifier was developed via the online-learning approach described in the RLSuite manuscript. The classifier evaluates features engineered from the RLFS Z score distribution, specifically, the following features:

feature_key <- dplyr::tribble(
  ~feature, ~description,
  "Z1", "mean of Z",
  "Z2", "variance of Z",
  "Zacf1", "mean of Z ACF",
  "Zacf2", "variance of Z ACF",
  "ReW1", "mean of FT of Z (real part)",
  "ReW2", "variance of FT of Z (real part)",
  "ImW1", "mean of FT of Z (imaginary part)",
  "ImW2", "variance of FT of Z (imaginary part)",
  "ReWacf1", "mean of FT of Z ACF (real part)",
  "ReWacf2", "variance of FT of Z ACF (real part)",
  "ImWacf1", "mean of FT of Z ACF (imaginary part)",
  "ImWacf2", "variance of FT of Z ACF (imaginary part)"
) 

dplyr::tibble(
  pred$Features
) %>% 
  dplyr::right_join(feature_key, by = "feature") %>%
  dplyr::relocate(description, .after = feature) %>%
  kableExtra::kable(caption = paste0(
    "Abbreviations: Z, Z-score",
    " distribution; ACF, autocorrelation function; FT, Fourier Transform."
  )) %>% 
  kableExtra::kable_material(c("striped", "hover"), 
                             position = "left", full_width=FALSE)

From these features, classification was performed to derive a prediction (predicted label) regarding whether the sample mapped R-loops or not. In short, "POS" indicates any sample for which all the following are true:

1. Criteria 1: The RLFS Permutation test P value is significant (p < .05)
2. Criteria 2: The Z-score distribution middle is > 0.
3. Criteria 3: The Z-score distribution middle is > the start and the end.
4. Criteria 4: The model predicts a label of "POS".

The criteria for r object@metadata$sampleName are shown below:

dplyr::tibble(
    Criteria = paste0(seq(4),". ", names(pred$Criteria)),
    Result = unlist(pred$Criteria)
) %>%
    kableExtra::kable(caption = paste0(
        "Results from quality analysis of <strong>",
        object@metadata$sampleName, "</strong>")
    ) %>%
    kableExtra::kable_material(c("striped", "hover"),
                               full_width=FALSE,
                               position="float_right") %>%
    kableExtra::column_spec(
        column = 2, 
        color = ifelse(unlist(pred$Criteria), "#A45BA4", "red"),
        bold = TRUE
    ) 

These results led to the final prediction: r verd.

For additional detail, please refer to the RLSeq::predictCondition documentation (link{target="_blank"}).

---

Bishop-Laboratory/RLSeq documentation built on Jan. 28, 2023, 11:38 p.m.