biopipen 0.29.0__py3-none-any.whl → 0.29.1__py3-none-any.whl

biopipen/__init__.py

@@ -1 +1 @@
-__version__ = "0.29.0"
+__version__ = "0.29.1"

biopipen/ns/plot.py

@@ -221,7 +221,7 @@ class Manhattan(Proc):
         "label_col": None,
         "devpars": {"res": 100, "width": 1000, "height": 500},
         "zoom_devpars": {"width": 500, "height": None, "res": None},
-        "title": "Manhattan Plot",
+        "title": None,
         "ylabel": "-log10(p-value)",
         "rescale": True,
         "rescale_ratio_threshold": 5,
@@ -245,6 +245,11 @@ class QQPlot(Proc):
         infile: The input file for data
             It should contain at least one column of p-values or the values to be
             plotted. Header is required.
+        theorfile: The file for theoretical values (optional)
+            This file should contain at least one column of theoretical values.
+            The values will be passed to `envs.theor_qfunc` to calculate the theoretical
+            quantiles.
+            Header is required.
 
     Output:
         outfile: The output figure file
@@ -266,33 +271,86 @@ class QQPlot(Proc):
         kind (choice): The kind of the plot, `qq` or `pp`
             - qq: QQ-plot
             - pp: PP-plot
-        band (ns): The arguments for `stat_qq_band()` or `stat_pp_band()`
+        theor_col: The column for theoretical values in `in.theorfile` if provided,
+            otherwise in `in.infile`.
+            An integer (1-based) or a string indicating the column name.
+            If `distribution` of `band`, `line`, or `point` is `custom`, this column
+            must be provided.
+        theor_trans: The transformation of the theoretical values.
+            The `theor_funs` have default functions to take the theoretical values.
+            This transformation will be applied to the theoretical values before
+            passing to the `theor_funs`.
+        theor_funs (ns): The R functions to generate density, quantile and deviates
+            of the theoretical distribution base on the theoretical values
+            if `distribution` of `band`, `line`, or `point` is `custom`.
+            - dcustom: The density function, used by band
+            - qcustom: The quantile function, used by point
+            - rcustom: The deviates function, used by line
+        args (ns): The common arguments for `envs.band`, `envs.line` and `envs.point`.
+            - distribution: The distribution of the theoretical quantiles
+                When `custom` is used, the `envs.theor_col` should be provided and
+                `values` will be added to `dparams` automatically.
+            - dparams (type=json): The parameters for the distribution
+            - <more>: Other shared arguments between `stat_*_band`, `stat_*_line`
+                and `stat_*_point`.
+        band (ns): The arguments for `stat_qq_band()` or `stat_pp_band()`.
             See <https://rdrr.io/cran/qqplotr/man/stat_qq_band.html> and
             <https://rdrr.io/cran/qqplotr/man/stat_pp_band.html>.
+            Set to `None` or `band.disabled` to True to disable the band.
+            - disabled (flag): Disable the band
+            - distribution: The distribution of the theoretical quantiles
+                When `custom` is used, the `envs.theor_col` should be provided and
+                `values` will be added to `dparams` automatically.
+            - dparams (type=json): The parameters for the distribution
             - <more>: Additional arguments for `stat_qq_band()` or `stat_pp_band()`
-        line (ns): The arguments for `stat_qq_line()` or `stat_pp_line()`
+        line (ns): The arguments for `stat_qq_line()` or `stat_pp_line()`.
             See <https://rdrr.io/cran/qqplot/man/stat_qq_line.html> and
             <https://rdrr.io/cran/qqplot/man/stat_pp_line.html>.
+            Set to `None` or `line.disabled` to True to disable the line.
+            - disabled (flag): Disable the line
+            - distribution: The distribution of the theoretical quantiles
+                When `custom` is used, the `envs.theor_col` should be provided and
+                `values` will be added to `dparams` automatically.
+            - dparams (type=json): The parameters for the distribution
             - <more>: Additional arguments for `stat_qq_line()` or `stat_pp_line()`
-        point (ns): The arguments for `geom_qq_point()` or `geom_pp_point()`
+        point (ns): The arguments for `geom_qq_point()` or `geom_pp_point()`.
             See <https://rdrr.io/cran/qqplot/man/stat_qq_point.html> and
             <https://rdrr.io/cran/qqplot/man/stat_pp_point.html>.
+            Set to `None` or `point.disabled` to True to disable the point.
+            - disabled (flag): Disable the point
+            - distribution: The distribution of the theoretical quantiles
+                When `custom` is used, the `envs.theor_col` should be provided and
+                `values` will be added to `dparams` automatically.
+            - dparams (type=json): The parameters for the distribution
+            - <more>: Additional arguments for `geom_qq_point()` or `geom_pp_point()`
         ggs (list): Additional ggplot expression to adjust the plot.
     """
-    input = "infile:file"
+    input = "infile:file, theorfile:file"
     output = "outfile:file:{{in.infile | stem}}.{{envs.kind}}.png"
     lang = config.lang.rscript
     envs = {
         "val_col": 1,
+        "theor_col": None,
+        "theor_trans": None,
+        "theor_funs": {
+            "dcustom": """
+              function(x, values, ...) {
+                density(values, from = min(values), to = max(values), n = length(x))$y
+              }
+            """,
+            "qcustom": "function(p, values, ...) {quantile(values, probs = p)}",
+            "rcustom": "function(n, values, ...) { sample(values, n, replace = TRUE) }",
+        },
+        "args": {"distribution": "norm", "dparams": {}},
         "devpars": {"res": 100, "width": 1000, "height": 1000},
         "xlabel": "Theoretical Quantiles",
         "ylabel": "Observed Quantiles",
         "title": "QQ-plot",
         "trans": None,
         "kind": "qq",
-        "band": {},
-        "line": {},
-        "point": {},
+        "band": {"disabled": False, "distribution": None, "dparams": None},
+        "line": {"disabled": False, "distribution": None, "dparams": None},
+        "point": {"disabled": False, "distribution": None, "dparams": None},
         "ggs": None,
     }
     script = "file://../scripts/plot/QQPlot.R"

biopipen/ns/{regulation.py → regulatory.py}

@@ -1,4 +1,4 @@
-"""Provides processes for the regulation related"""
+"""Provides processes for the regulatory related"""
 
 from ..core.proc import Proc
 from ..core.config import config
@@ -86,7 +86,7 @@ class MotifScan(Proc):
         "q_cutoff": False,
         "args": {},
     }
-    script = "file://../scripts/regulation/MotifScan.py"
+    script = "file://../scripts/regulatory/MotifScan.py"
 
 
 class MotifAffinityTest(Proc):
@@ -211,4 +211,4 @@ class MotifAffinityTest(Proc):
         "motifbreakr_args": {"method": "default"},
         "atsnp_args": {"padj_cutoff": True, "padj": "BH", "p": "pval_diff"},
     }
-    script = "file://../scripts/regulation/MotifAffinityTest.R"
+    script = "file://../scripts/regulatory/MotifAffinityTest.R"

biopipen/ns/scrna.py

@@ -53,7 +53,7 @@ class SeuratPreparing(Proc):
 
     See also
     - <https://satijalab.org/seurat/articles/pbmc3k_tutorial.html#standard-pre-processing-workflow-1)>
-    - <https://nbisweden.github.io/workshop-scRNAseq/labs/compiled/seurat/seurat_01_qc.html#Create_one_merged_object>
+    - <https://satijalab.org/seurat/articles/integration_introduction>
 
     This process will read the scRNA-seq data, based on the information provided by
     `SampleInfo`, specifically, the paths specified by the `RNAData` column.
@@ -210,6 +210,19 @@ class SeuratPreparing(Proc):
             - PCs (type=int): Number of PCs to use for 'doubletFinder' function.
             - doublets (type=float): Number of expected doublets as a proportion of the pool size.
             - pN (type=float): Number of doublets to simulate as a proportion of the pool size.
+            - ncores (type=int): Number of cores to use for `DoubletFinder::paramSweep`.
+                Set to `None` to use `envs.ncores`.
+                Since parallelization of the function usually exhausts memory, if big `envs.ncores` does not work
+                for `DoubletFinder`, set this to a smaller number.
+
+        cache (type=auto): Whether to cache the information at different steps.
+            If `True`, the seurat object will be cached in the job output directory, which will be not cleaned up when job is rerunning.
+            The cached seurat object will be saved as `<signature>.<kind>.RDS` file, where `<signature>` is the signature determined by
+            the input and envs of the process.
+            See <https://github.com/satijalab/seurat/issues/7849>, <https://github.com/satijalab/seurat/issues/5358> and
+            <https://github.com/satijalab/seurat/issues/6748> for more details also about reproducibility issues.
+            To not use the cached seurat object, you can either set `cache` to `False` or delete the cached file at
+            `<signature>.RDS` in the cache directory.
 
     Requires:
         r-seurat:
@@ -238,7 +251,8 @@ class SeuratPreparing(Proc):
             "min_cells": 5,
         },
         "IntegrateLayers": {"method": "harmony"},
-        "DoubletFinder": {"PCs": 0, "pN": 0.25, "doublets": 0.075},
+        "DoubletFinder": {"PCs": 0, "pN": 0.25, "doublets": 0.075, "ncores": 1},
+        "cache": config.path.tmpdir,
     }
     script = "file://../scripts/scrna/SeuratPreparing.R"
     plugin_opts = {

biopipen/ns/stats.py

@@ -73,11 +73,103 @@ class ChowTest(Proc):
     script = "file://../scripts/stats/ChowTest.R"
 
 
+class Mediation(Proc):
+    """Mediation analysis.
+
+    The flowchart of mediation analysis:
+
+    ![Mediation Analysis](https://library.virginia.edu/sites/default/files/inline-images/mediation_flowchart-1.png)
+
+    Reference:
+        - <https://library.virginia.edu/data/articles/introduction-to-mediation-analysis>
+        - <https://en.wikipedia.org/wiki/Mediation_(statistics)>
+        - <https://tilburgsciencehub.com/topics/analyze/regression/linear-regression/mediation-analysis/>
+        - <https://ademos.people.uic.edu/Chapter14.html>
+
+    Input:
+        infile: The input data file. The rows are samples and the columns are
+            features. It must be tab-delimited.
+            ```
+            Sample   F1   F2   F3   ...   Fn
+            S1       1.2  3.4  5.6        7.8
+            S2       2.3  4.5  6.7        8.9
+            ...
+            Sm       5.6  7.8  9.0        1.2
+            ```
+        fmlfile: The formula file.
+            ```
+            Case   M   Y   X   Cov     Model_M    Model_Y
+            Case1  F1  F2  F3  F4,F5   glm        lm
+            ...
+            ```
+            Where Y is the outcome variable, X is the predictor variable, M is the
+            mediator variable, and Case is the case name. Model_M and Model_Y are the
+            models for M and Y, respectively.
+            `envs.cases` will be ignored if this is provided.
+
+    Output:
+        outfile: The output file.
+            Columns to help understand the results:
+            Total Effect: a total effect of X on Y (without M) (`Y ~ X`).
+            ADE: A Direct Effect of X on Y after taking into account a mediation effect of M (`Y ~ X + M`).
+            ACME: The Mediation Effect, the total effect minus the direct effect,
+            which equals to a product of a coefficient of X in the second step and a coefficient of M in the last step.
+            The goal of mediation analysis is to obtain this indirect effect and see if it's statistically significant.
+
+    Envs:
+        ncores (type=int): Number of cores to use for parallelization for cases.
+        sims (type=int): Number of Monte Carlo draws for nonparametric bootstrap or quasi-Bayesian approximation.
+            Will be passed to `mediation::mediate` function.
+        args (ns): Other arguments passed to `mediation::mediate` function.
+            - <more>: More arguments passed to `mediation::mediate` function.
+                See: <https://rdrr.io/cran/mediation/man/mediate.html>
+        padj (choice): The method for (ACME) p-value adjustment.
+            - none: No p-value adjustment (no Padj column in outfile).
+            - holm: Holm-Bonferroni method.
+            - hochberg: Hochberg method.
+            - hommel: Hommel method.
+            - bonferroni: Bonferroni method.
+            - BH: Benjamini-Hochberg method.
+            - BY: Benjamini-Yekutieli method.
+            - fdr: FDR correction method.
+        cases (type=json): The cases for mediation analysis.
+            Ignored if `in.fmlfile` is provided.
+            A json/dict with case names as keys and values as a dict of M, Y, X, Cov, Model_M, Model_Y.
+            For example:
+            ```json
+            {
+                "Case1": {
+                    "M": "F1",
+                    "Y": "F2",
+                    "X": "F3",
+                    "Cov": "F4,F5",
+                    "Model_M": "glm",
+                    "Model_Y": "lm"
+                },
+                ...
+            }
+            ```
+        transpose_input (flag): Whether to transpose the input file.
+    """  # noqa: E501
+    input = "infile:file, fmlfile:file"
+    output = "outfile:file:{{in.infile | stem}}.mediation.txt"
+    lang = config.lang.rscript
+    envs = {
+        "ncores": config.misc.ncores,
+        "sims": 1000,
+        "args": {},
+        "padj": "none",
+        "cases": {},
+        "transpose_input": False,
+    }
+    script = "file://../scripts/stats/Mediation.R"
+
+
 class LiquidAssoc(Proc):
     """Liquid association tests.
 
     See Also https://github.com/gundt/fastLiquidAssociation
-    Requieres https://github.com/pwwang/fastLiquidAssociation
+    Requires https://github.com/pwwang/fastLiquidAssociation
 
     Input:
         infile: The input data file. The rows are samples and the columns are

biopipen/scripts/delim/SampleInfo.R

@@ -88,7 +88,11 @@ for (name in names(stats)) {
     group <- if (is.null(stat$group)) sym("..group") else sym(stat$group)
     count_on <- paste0("..count.", stat$on)
     if (!is_continuous) {
-        data <- data %>% add_count(!!group, name = count_on)
+        if (!is.null(stat$each)) {
+            data <- data %>% add_count(!!group, !!sym(stat$each), name = count_on)
+        } else {
+            data <- data %>% add_count(!!group, name = count_on)
+        }
     }
 
     if (is.null(stat$devpars)) {
@@ -141,18 +145,19 @@ for (name in names(stats)) {
         } else {
             data <- data %>%
                 distinct(!!group, !!sym(stat$each), .keep_all = TRUE) %>%
+                mutate(!!group := factor(!!group, levels = unique(!!group))) %>%
                 group_by(!!sym(stat$each))
         }
         p <- ggplot(
-            data %>% arrange(!!group),
-            aes(x = "", y = !!sym(count_on), fill = !!group, label = !!sym(count_on))
+            data %>% mutate(.size = sum(!!sym(count_on))),
+            aes(x = sqrt(.size) / 2, width = sqrt(.size), y = !!sym(count_on), fill = !!group, label = !!sym(count_on))
         ) +
-            geom_bar(stat="identity", width=1, color="white", position = position_stack(reverse = TRUE)) +
+            geom_bar(stat="identity", color="white", position = position_fill(reverse = TRUE)) +
             coord_polar("y", start = 0) +
             theme_void() +
             theme(plot.title = element_text(hjust = 0.5)) +
             geom_label_repel(
-                position = position_stack(vjust = 0.5),
+                position = position_fill(reverse = TRUE,vjust = .5),
                 color="#333333",
                 fill="#EEEEEE",
                 size=4

biopipen/scripts/plot/Manhattan.R

@@ -105,6 +105,7 @@ args$signif <- signif
 args$plot.title <- title
 args$rescale <- rescale
 args$rescale.ratio.threshold <- rescale_ratio_threshold
+args$y.label <- ylabel
 if (!is.null(hicolors)) { args$color.by.highlight <- TRUE }
 if (!is.null(label_col)) { args$label.colname <- ".label" }
 g <- do_call(manhattan_plot, args)
@@ -114,10 +115,15 @@ print(g)
 dev.off()
 
 # zoom into chromosomes
+all_chroms <- as.character(unique(mpdata$data[[mpdata$chr.colname]]))
 if (!is.null(zoom)) {
     log_info("Zooming into chromosomes ...")
     zoom <- norm_chroms(zoom)
     for (z in zoom) {
+        if (!z %in% all_chroms) {
+            log_warn("- {z}: not found in data")
+            next
+        }
         log_info("- {z}")
         args_z <- args
         args_z$chromosome <- z

biopipen/scripts/plot/QQPlot.R

@@ -1,5 +1,7 @@
 source("{{biopipen_dir}}/utils/misc.R")
 
+library(rlang)
+library(stats)
 library(ggplot2)
 library(ggprism)
 library(qqplotr)
@@ -7,50 +9,132 @@ library(qqplotr)
 theme_set(theme_prism())
 
 infile <- {{in.infile | r}}
+theorfile <- {{in.theorfile | r}}
 outfile <- {{out.outfile | r}}
 val_col <- {{envs.val_col | r}}
+theor_col <- {{envs.theor_col | r}}
+theor_trans <- {{envs.theor_trans | r}}
+theor_funs <- {{envs.theor_funs | r}}
 devpars <- {{envs.devpars | r}}
 title <- {{envs.title | r}}
 xlabel <- {{envs.xlabel | r}}
 ylabel <- {{envs.ylabel | r}}
 kind <- {{envs.kind | r}}
 trans <- {{envs.trans | r}}
+args <- {{envs.args | r}}
 band_args <- {{envs.band | r}}
 line_args <- {{envs.line | r}}
 point_args <- {{envs.point | r}}
 ggs <- {{envs.ggs | r}}
 
+.eval_fun <- function(fun) {
+    if (is.character(fun)) {
+        fun <- trimws(fun)
+        if (grepl("^-\\s*[a-zA-Z\\.][0-9a-zA-Z\\._]*$", fun)) {
+            fun <- trimws(substring(fun, 2))
+            fun <- eval(parse(text = fun))
+            return(function(x) -fun(x))
+        } else {
+            return(eval(parse(text = fun)))
+        }
+    } else {
+        return(fun)
+    }
+}
+
 indata <- read.table(infile, header=TRUE, sep="\t", stringsAsFactors=FALSE, check.names = FALSE)
-if (is.numeric(val_col)) { val_col <- colnames(indata)[val_col] }
+if (is.numeric(val_col)) {
+    val_col <- colnames(indata)[val_col]
+}
+if (!is.null(trans)) {
+    trans <- .eval_fun(trans)
+    indata[[val_col]] <- trans(indata[[val_col]])
+}
+
+if (!is.null(theor_col)) {
+    if (is.numeric(theor_col)) {
+        theor_col <- colnames(theor)[theor_col]
+    }
+
+    if (!is.null(theorfile)) {
+        theor <- read.table(theorfile, header=TRUE, sep="\t", stringsAsFactors=FALSE, check.names = FALSE)
+        theor_vals <- theor[[theor_col]]
+    } else {
+        theor_vals <- indata[[theor_col]]
+    }
+
+    if (!is.null(theor_trans)) {
+        theor_trans <- .eval_fun(theor_trans)
+        theor_vals <- theor_trans(theor_vals)
+    }
+    theor_vals <- sort(na.omit(theor_vals))
+}
 
 band_fun <- ifelse(kind == "pp", stat_pp_band, stat_qq_band)
 line_fun <- ifelse(kind == "pp", stat_pp_line, stat_qq_line)
 point_fun <- ifelse(kind == "pp", stat_pp_point, stat_qq_point)
 
-title <- title %||% waiver()
-xlabel <- xlabel %||% waiver()
-ylabel <- ylabel %||% waiver()
+for (fun in names(theor_funs)) {
+    assign(fun, .eval_fun(theor_funs[[fun]]))
+}
 
-if (!is.null(trans)) {
-    trans <- trimws(trans)
-    if (trans == "-log10") {
-        trans <- function(x) -log10(x)
+if (!is.null(band_args) || isFALSE(band_args)) {
+    if (isTRUE(band_args$disabled)) {
+        band_args <- NULL
     } else {
-        trans <- eval(parse(text = trans))
+        band_args$disabled <- NULL
+        band_args <- list_update(band_args, args)
+        if (band_args$distribution == "custom") {
+            band_args$dparams <- band_args$dparams %||% list()
+            band_args$dparams$values <- theor_vals
+        }
+    }
+}
+if (!is.null(line_args) || isFALSE(line_args)) {
+    if (isTRUE(line_args$disabled)) {
+        line_args <- NULL
+    } else {
+        line_args$disabled <- NULL
+        line_args <- list_update(line_args, args)
+        if (line_args$distribution == "custom") {
+            line_args$dparams <- line_args$dparams %||% list()
+            line_args$dparams$values <- theor_vals
+        }
+    }
+}
+if (!is.null(point_args) || isFALSE(point_args)) {
+    if (isTRUE(point_args$disabled)) {
+        point_args <- NULL
+    } else {
+        point_args$disabled <- NULL
+        point_args <- list_update(point_args, args)
+        if (point_args$distribution == "custom") {
+            point_args$dparams <- point_args$dparams %||% list()
+            point_args$dparams$values <- theor_vals
+        }
     }
-
-    indata$.trans_val <- trans(indata[[val_col]])
-    val_col <- ".trans_val"
 }
 
-indata <- indata[!is.na(indata[[val_col]]), , drop=FALSE]
+title <- title %||% waiver()
+xlabel <- xlabel %||% waiver()
+ylabel <- ylabel %||% waiver()
+
+indata <- indata[complete.cases(indata), , drop = FALSE]
+indata <- indata[order(indata[[val_col]]), , drop = FALSE]
 
 p <- ggplot(data = indata, mapping = aes(sample = !!sym(val_col))) +
-    do_call(band_fun, band_args) +
-    do_call(line_fun, line_args) +
-    do_call(point_fun, point_args) +
     labs(title = title, x = xlabel, y = ylabel)
 
+if (!is.null(band_args)) {
+    p <- p + do_call(band_fun, band_args)
+}
+if (!is.null(line_args)) {
+    p <- p + do_call(line_fun, line_args)
+}
+if (!is.null(point_args)) {
+    p <- p + do_call(point_fun, point_args)
+}
+
 if (!is.null(ggs)) {
     for (gg in ggs) {
         p <- p + eval(parse(text = gg))

biopipen/scripts/{regulation → regulatory}/MotifAffinityTest.R

@@ -1,4 +1,4 @@
-# Script for regulation.MotifAffinityTest
+# Script for regulatory.MotifAffinityTest
 
 source("{{biopipen_dir}}/utils/misc.R")
 library(BiocParallel)
@@ -215,12 +215,12 @@ tool <- match.arg(tool, c("motifbreakr", "atsnp"))
 
 if (tool == "motifbreakr") {
     motifbreakr_args <- {{envs.motifbreakr_args | r}}
-    {% set sourcefile = biopipen_dir | joinpaths: "scripts", "regulation", "MotifAffinityTest_MotifBreakR.R" %}
+    {% set sourcefile = biopipen_dir | joinpaths: "scripts", "regulatory", "MotifAffinityTest_MotifBreakR.R" %}
     # {{ sourcefile | getmtime }}
     source("{{sourcefile}}")
 } else {  # atsnp
     atsnp_args <- {{envs.atsnp_args | r}}
-    {% set sourcefile = biopipen_dir | joinpaths: "scripts", "regulation", "MotifAffinityTest_AtSNP.R" %}
+    {% set sourcefile = biopipen_dir | joinpaths: "scripts", "regulatory", "MotifAffinityTest_AtSNP.R" %}
     # {{ sourcefile | getmtime }}
     source("{{sourcefile}}")
 }

biopipen/scripts/{regulation → regulatory}/MotifScan.py

@@ -1,4 +1,4 @@
-"""Script for regulation.MotifScan"""
+"""Script for regulatory.MotifScan"""
 import re
 
 # Paths may be passed in args or to motifdb

biopipen/scripts/scrna/MarkersFinder.R

@@ -65,6 +65,19 @@ if (ncores > 1) {
 log_info("- Reading Seurat object ...")
 srtobj <- readRDS(srtfile)
 defassay <- DefaultAssay(srtobj)
+if (defassay == "SCT" && !"PrepSCTFindMarkers" %in% names(srtobj@commands)) {
+    log_warn("  SCTransform used but PrepSCTFindMarkers not applied, running ...")
+
+    srtobj <- PrepSCTFindMarkers(srtobj)
+    # compose a new SeuratCommand to record it to srtobj@commands
+    scommand <- srtobj@commands$FindClusters
+    scommand@name <- "PrepSCTFindMarkers"
+    scommand@time.stamp <- Sys.time()
+    scommand@assay.used <- "SCT"
+    scommand@call.string <- "PrepSCTFindMarkers(object = srtobj)"
+    scommand@params <- list()
+    srtobj@commands$PrepSCTFindMarkers <- scommand
+}
 
 if (!is.null(mutaters) && length(mutaters) > 0) {
     log_info("- Mutating meta data ...")
@@ -472,33 +485,30 @@ find_markers <- function(findmarkers_args, find_all = FALSE) {
             p_val_adj = numeric()
         )
     }
+
+    call_findmarkers <- function(fn, args) {
+        if (find_all) {
+            do_call(fn, args)
+        } else {
+            do_call(fn, args) %>% rownames_to_column("gene")
+        }
+    }
     markers <- tryCatch({
-        do_call(fun, findmarkers_args) %>% rownames_to_column("gene")
+        call_findmarkers(fun, findmarkers_args)
     }, error = function(e) {
-        # Object contains multiple models with unequal library sizes.
-        # Run `PrepSCTFindMarkers()` before running `FindMarkers()`.
-        if (grepl("PrepSCTFindMarkers", e$message)) {
-            log_warn("  Running PrepSCTFindMarkers ...")
-            findmarkers_args$object <<- PrepSCTFindMarkers(findmarkers_args$object)
-            tryCatch({
-                do_call(fun, findmarkers_args) %>% rownames_to_column("gene")
-            }, error = function(err) {
-                log_warn(paste0("  ", err$message))
-                empty
-            })
-        } else {
-            log_warn(paste0("  ", e$message))
-            empty
+        if (!grepl("PrepSCTFindMarkers", e$message) && defassay == "SCT") {
+            log_warn(paste0("  ! ", e$message))
         }
+        empty
     })
 
     if (nrow(markers) == 0 && defassay == "SCT") {
-        log_warn("  No markers found from SCT assay, trying recorrect_umi = FALSE")
+        log_warn("  ! No markers found from SCT assay, trying recorrect_umi = FALSE")
         findmarkers_args$recorrect_umi <- FALSE
         markers <- tryCatch({
-            do_call(fun, findmarkers_args) %>% rownames_to_column("gene")
+            call_findmarkers(fun, findmarkers_args)
         }, error = function(e) {
-            log_warn(paste0("  ", e$message))
+            log_warn(paste0("  ! ", e$message))
             empty
         })
     }

biopipen/scripts/scrna/SeuratClustering.R

@@ -202,6 +202,14 @@ if (DefaultAssay(sobj) == "SCT") {
         # https://github.com/satijalab/seurat/issues/6968
     log_info("Running PrepSCTFindMarkers ...")
     sobj <- PrepSCTFindMarkers(sobj)
+    # compose a new SeuratCommand to record it to sobj@commands
+    scommand <- sobj@commands$FindClusters
+    scommand@name <- "PrepSCTFindMarkers"
+    scommand@time.stamp <- Sys.time()
+    scommand@assay.used <- "SCT"
+    scommand@call.string <- "PrepSCTFindMarkers(object = sobj)"
+    scommand@params <- list()
+    sobj@commands$PrepSCTFindMarkers <- scommand
 }
 
 log_info("Saving results ...")