From 1505e43ddf49f918dbd7d11e30d392754dc8e632 Mon Sep 17 00:00:00 2001
From: Tony Wu <wu.anthon@northeastern.edu>
Date: Mon, 6 Jul 2026 14:55:29 -0400
Subject: [PATCH] fix(turnover): Fix DIANN 1.8 processing of protein turnover
 labeling

---
 R/clean_DIANN.R                     | 19 +++++++++++--------
 R/converters_DIANNtoMSstatsFormat.R | 12 +++++++-----
 inst/tinytest/test_clean_DIANN.R    | 15 +++++++++++++++
 man/DIANNtoMSstatsFormat.Rd         | 12 +++++++-----
 man/MSstatsClean.Rd                 | 12 +++++++-----
 man/dot-cleanRawDIANN.Rd            | 12 +++++++-----
 6 files changed, 54 insertions(+), 28 deletions(-)

diff --git a/R/clean_DIANN.R b/R/clean_DIANN.R
index 130066a9..5eb0b0cf 100644
--- a/R/clean_DIANN.R
+++ b/R/clean_DIANN.R
@@ -232,10 +232,11 @@
 #'
 #' \strong{ModifiedSequence-parsing path} (\code{has_channel = FALSE}):
 #' \code{PeptideSequence} (the retained \code{ModifiedSequence}) is matched
-#' against SILAC suffixes of the form \code{(SILAC-<AA>-H)} or
-#' \code{(SILAC-<AA>-L)}, where \code{<AA>} is any code in
-#' \code{labeledAminoAcids}.  Sequences carrying neither suffix receive
-#' \code{IsotopeLabelType = NA}.  The SILAC suffix is stripped from
+#' against label suffixes of the form \code{(<label>-<AA>-H)} or
+#' \code{(<label>-<AA>-L)}, where \code{<AA>} is any code in
+#' \code{labeledAminoAcids} and \code{<label>} is any non-dash token (e.g.
+#' \code{SILAC} or \code{label}).  Sequences carrying neither suffix receive
+#' \code{IsotopeLabelType = NA}.  The label suffix is stripped from
 #' \code{PeptideSequence} after classification.
 #'
 #' @param dn_input \code{data.table} after column renaming.
@@ -265,9 +266,11 @@
         dn_input[, Channel := NULL]
     } else {
         aa_pattern <- paste0(labeledAminoAcids, collapse = "|")
-        heavy_regex <- paste0("\\(SILAC-(?:", aa_pattern, ")-H\\)")
-        light_regex <- paste0("\\(SILAC-(?:", aa_pattern, ")-L\\)")
-        strip_regex <- paste0("\\(SILAC-(?:", aa_pattern, ")-[HL]\\)")
+        # The label token after the first parenthesis is not always "SILAC"
+        # (e.g. "(SILAC-L-L)" vs "(label-L-L)"), so match any non-dash token there.
+        heavy_regex <- paste0("\\([^-]+-(?:", aa_pattern, ")-H\\)")
+        light_regex <- paste0("\\([^-]+-(?:", aa_pattern, ")-L\\)")
+        strip_regex <- paste0("\\([^-]+-(?:", aa_pattern, ")-[HL]\\)")
 
         dn_input <- .classifyIsotopeLabelType(dn_input, heavy_regex, light_regex)
         dn_input[, PeptideSequence := gsub(strip_regex, "", PeptideSequence, perl = TRUE)]
@@ -276,7 +279,7 @@
     if (all(is.na(dn_input[["IsotopeLabelType"]]))) {
         warning("labeledAminoAcids was provided but no rows were classified as H or L. ",
                 "Check that the input contains either a Channel column with H/L values ",
-                "or ModifiedSequence entries with (SILAC-<AA>-H)/(SILAC-<AA>-L) suffixes.")
+                "or ModifiedSequence entries with (<label>-<AA>-H)/(<label>-<AA>-L) suffixes.")
     }
 
     dn_input
diff --git a/R/converters_DIANNtoMSstatsFormat.R b/R/converters_DIANNtoMSstatsFormat.R
index 78f3f163..7b291163 100644
--- a/R/converters_DIANNtoMSstatsFormat.R
+++ b/R/converters_DIANNtoMSstatsFormat.R
@@ -36,11 +36,13 @@
 #' \strong{ModifiedSequence-parsing path} (DIA-NN 1.x exports without a
 #' \code{Channel} column): when \code{labeledAminoAcids} is non-\code{NULL}
 #' and no \code{Channel} column is present, each \code{ModifiedSequence} is
-#' inspected for SILAC suffixes of the form \code{(SILAC-<AA>-H)} or
-#' \code{(SILAC-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
-#' codes.  Matching sequences are classified as \code{"H"} or \code{"L"};
-#' sequences carrying neither suffix receive \code{IsotopeLabelType = NA}.
-#' The SILAC suffix is then stripped from \code{PeptideSequence}.
+#' inspected for label suffixes of the form \code{(<label>-<AA>-H)} or
+#' \code{(<label>-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
+#' codes and \code{<label>} is any non-dash token (e.g. \code{SILAC} or
+#' \code{label}).  Matching sequences are classified as \code{"H"} or
+#' \code{"L"}; sequences carrying neither suffix receive
+#' \code{IsotopeLabelType = NA}.  The label suffix is then stripped from
+#' \code{PeptideSequence}.
 #'
 #' When \code{NULL} (default), protein-turnover mode is disabled and all
 #' peptides receive \code{IsotopeLabelType = "Light"}.
diff --git a/inst/tinytest/test_clean_DIANN.R b/inst/tinytest/test_clean_DIANN.R
index 9c69fcd8..6d46dfbb 100644
--- a/inst/tinytest/test_clean_DIANN.R
+++ b/inst/tinytest/test_clean_DIANN.R
@@ -76,3 +76,18 @@ result_null = MSstatsConvert:::.assignDIANNIsotopeLabelType(
 )
 expect_false("IsotopeLabelType" %in% colnames(result_null))
 expect_equal(nrow(result_null), 2L)
+
+# ModifiedSequence-parsing path: classification and stripping are agnostic to
+# the label token after the first parenthesis. The suffix may be (SILAC-<AA>-H)
+# or any (<label>-<AA>-H), e.g. (label-K-H).
+dt_label = data.table::data.table(
+    PeptideSequence = c("PEPTIDEK(SILAC-K-H)", "PEPTIDEK(SILAC-K-L)",
+                        "PEPTIDEK(label-K-H)", "PEPTIDEK(label-K-L)",
+                        "PEPTIDEK")
+)
+result_label = MSstatsConvert:::.assignDIANNIsotopeLabelType(
+    dt_label, labeledAminoAcids = c("K"), has_channel = FALSE
+)
+expect_equal(result_label$IsotopeLabelType, c("H", "L", "H", "L", NA_character_))
+# Suffix is stripped regardless of the label token
+expect_equal(unique(result_label$PeptideSequence), "PEPTIDEK")
diff --git a/man/DIANNtoMSstatsFormat.Rd b/man/DIANNtoMSstatsFormat.Rd
index e89f9cdb..de76501c 100644
--- a/man/DIANNtoMSstatsFormat.Rd
+++ b/man/DIANNtoMSstatsFormat.Rd
@@ -79,11 +79,13 @@ in \code{labeledAminoAcids} are \strong{not} used to validate or filter
 \strong{ModifiedSequence-parsing path} (DIA-NN 1.x exports without a
 \code{Channel} column): when \code{labeledAminoAcids} is non-\code{NULL}
 and no \code{Channel} column is present, each \code{ModifiedSequence} is
-inspected for SILAC suffixes of the form \code{(SILAC-<AA>-H)} or
-\code{(SILAC-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
-codes.  Matching sequences are classified as \code{"H"} or \code{"L"};
-sequences carrying neither suffix receive \code{IsotopeLabelType = NA}.
-The SILAC suffix is then stripped from \code{PeptideSequence}.
+inspected for label suffixes of the form \code{(<label>-<AA>-H)} or
+\code{(<label>-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
+codes and \code{<label>} is any non-dash token (e.g. \code{SILAC} or
+\code{label}).  Matching sequences are classified as \code{"H"} or
+\code{"L"}; sequences carrying neither suffix receive
+\code{IsotopeLabelType = NA}.  The label suffix is then stripped from
+\code{PeptideSequence}.
 
 When \code{NULL} (default), protein-turnover mode is disabled and all
 peptides receive \code{IsotopeLabelType = "Light"}.}
diff --git a/man/MSstatsClean.Rd b/man/MSstatsClean.Rd
index bfa79d6a..1308147f 100644
--- a/man/MSstatsClean.Rd
+++ b/man/MSstatsClean.Rd
@@ -192,11 +192,13 @@ in \code{labeledAminoAcids} are \strong{not} used to validate or filter
 \strong{ModifiedSequence-parsing path} (DIA-NN 1.x exports without a
 \code{Channel} column): when \code{labeledAminoAcids} is non-\code{NULL}
 and no \code{Channel} column is present, each \code{ModifiedSequence} is
-inspected for SILAC suffixes of the form \code{(SILAC-<AA>-H)} or
-\code{(SILAC-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
-codes.  Matching sequences are classified as \code{"H"} or \code{"L"};
-sequences carrying neither suffix receive \code{IsotopeLabelType = NA}.
-The SILAC suffix is then stripped from \code{PeptideSequence}.
+inspected for label suffixes of the form \code{(<label>-<AA>-H)} or
+\code{(<label>-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
+codes and \code{<label>} is any non-dash token (e.g. \code{SILAC} or
+\code{label}).  Matching sequences are classified as \code{"H"} or
+\code{"L"}; sequences carrying neither suffix receive
+\code{IsotopeLabelType = NA}.  The label suffix is then stripped from
+\code{PeptideSequence}.
 
 When \code{NULL} (default), protein-turnover mode is disabled and all
 peptides receive \code{IsotopeLabelType = "Light"}.}
diff --git a/man/dot-cleanRawDIANN.Rd b/man/dot-cleanRawDIANN.Rd
index da61ebc1..b93fbe8d 100644
--- a/man/dot-cleanRawDIANN.Rd
+++ b/man/dot-cleanRawDIANN.Rd
@@ -59,11 +59,13 @@ in \code{labeledAminoAcids} are \strong{not} used to validate or filter
 \strong{ModifiedSequence-parsing path} (DIA-NN 1.x exports without a
 \code{Channel} column): when \code{labeledAminoAcids} is non-\code{NULL}
 and no \code{Channel} column is present, each \code{ModifiedSequence} is
-inspected for SILAC suffixes of the form \code{(SILAC-<AA>-H)} or
-\code{(SILAC-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
-codes.  Matching sequences are classified as \code{"H"} or \code{"L"};
-sequences carrying neither suffix receive \code{IsotopeLabelType = NA}.
-The SILAC suffix is then stripped from \code{PeptideSequence}.
+inspected for label suffixes of the form \code{(<label>-<AA>-H)} or
+\code{(<label>-<AA>-L)}, where \code{<AA>} is one of the supplied amino acid
+codes and \code{<label>} is any non-dash token (e.g. \code{SILAC} or
+\code{label}).  Matching sequences are classified as \code{"H"} or
+\code{"L"}; sequences carrying neither suffix receive
+\code{IsotopeLabelType = NA}.  The label suffix is then stripped from
+\code{PeptideSequence}.
 
 When \code{NULL} (default), protein-turnover mode is disabled and all
 peptides receive \code{IsotopeLabelType = "Light"}.}