docs(mTNBC): rename tnbc -> mTNBC and update plan.md with full QSP mo…

README.md

@@ -13,7 +13,7 @@
 | 참여자 | GitHub ID | 연구 주제 | 참고 논문 |
 |--------|-----------|-----------|-----------|
 | 박광원 | pangpangwon | Elagolix 임상 3상 데이터 기반 칼슘 항상성/골대사 QSP 모델 검증 | [Validation of a QSP model of calcium homeostasis](https://pubmed.ncbi.nlm.nih.gov/33963686/) |
-| 박정민 | wjdals011106 | 전이성 삼중음성유방암(mTNBC) 면역항암제 반응 예측 QSP 모델 | [A transcriptome-informed QSP model of mTNBC](https://www.science.org/doi/10.1126/sciadv.adg0289) |
+| 박정민 | wjdals011106 | 전이성 삼중음성유방암(mTNBC) 면역항암제 반응 예측 QSP 모델 | [A transcriptome-informed QSP model of mTNBC](https://www.science.org/doi/10.1126/sciadv.adg0289) · [🚀 Shiny App](https://wjdals011106.shinyapps.io/tnbc-qsp-simulator/) |
 | 배정현 | john83072 | 알츠하이머병 Aβ 플라크 표적 치료제 임상실패 원인 분석 QSP 모델 | [In silico analysis for reducing Aβ plaque](https://pubmed.ncbi.nlm.nih.gov/33938131/) |
 | 한상하 | magen2001 | 조현병 항정신병약 반응 예측 QSP 모델 | [Computer-based mechanistic schizophrenia disease model](https://doi.org/10.1371/journal.pone.0049732) |
 
@@ -213,6 +213,7 @@ capstone-qsp/
 
 ### 예시 앱
 - [Merigolix QSP Dashboard](https://pipetqsp.shinyapps.io/merigolix/)
+- [TNBC QSP Simulator (박정민)](https://wjdals011106.shinyapps.io/tnbc-qsp-simulator/)
 
 ---
 

mTNBC/R/03_events.R

@@ -0,0 +1,184 @@
+# =============================================================================
+# 03_events.R
+# TNBC QSP Model - Discrete Events
+# - Pembrolizumab dosing (200 mg IV q3w)
+# - Metastatic tumor seeding (at delay_Ln1, delay_Ln2, delay_other)
+# - Primary tumor surgery (at surgery_time, if do_surgery=1)
+# =============================================================================
+
+# -----------------------------------------------------------------------------
+# build_event_data()
+# Returns a data.frame of events for deSolve::lsoda (eventfun or event$data)
+# -----------------------------------------------------------------------------
+build_event_data <- function(p, t_start = 0, t_end = 700) {
+
+  events <- list()
+
+  # ---------------------------------------------------------------------------
+  # 1. PEMBROLIZUMAB DOSING (200 mg IV q3w)
+  # ---------------------------------------------------------------------------
+  # Convert dose to nM in central compartment (V_C = 5 L)
+  # 200 mg / 146700 g/mol = 1.3633e-3 mmol = 1.3633e-6 mol
+  # [nM] = mol / V_C(L) * 1e9 = 1.3633e-6 / 5 * 1e9 = 272.65 nM
+  dose_nM <- (p$dose_pembrolizumab / p$MW_pembrolizumab * 1e3) / p$V_C * 1e9  # nM
+
+  dose_times <- seq(t_start, t_end, by = p$dose_interval)
+  dose_times <- dose_times[dose_times <= t_end]
+
+  for (td in dose_times) {
+    events[[ length(events) + 1 ]] <- data.frame(
+      var   = "aPD1_C",
+      time  = td,
+      value = dose_nM,
+      method = "add"
+    )
+  }
+
+  # ---------------------------------------------------------------------------
+  # 2. METASTATIC TUMOR SEEDING
+  # ---------------------------------------------------------------------------
+  # At seeding time, set start flag and add initial cancer cells
+  # (start_Ln1, start_Ln2, start_other switch to 1)
+  seeding_vars <- list(
+    list(delay = p$delay_Ln1,   flag = "start_Ln1",   clones = c("C1_Ln1","C2_Ln1","C3_Ln1","C4_Ln1","C5_Ln1"),
+         ncells = c(p$ncells_C1_Ln1, p$ncells_C2_Ln1, p$ncells_C3_Ln1, p$ncells_C4_Ln1, p$ncells_C5_Ln1)),
+    list(delay = p$delay_Ln2,   flag = "start_Ln2",   clones = c("C1_Ln2","C2_Ln2","C3_Ln2","C4_Ln2","C5_Ln2"),
+         ncells = c(p$ncells_C1_Ln2, p$ncells_C2_Ln2, p$ncells_C3_Ln2, p$ncells_C4_Ln2, p$ncells_C5_Ln2)),
+    list(delay = p$delay_other, flag = "start_other", clones = c("C1_oth","C2_oth","C3_oth","C4_oth","C5_oth"),
+         ncells = c(p$ncells_C1_other, p$ncells_C2_other, p$ncells_C3_other, p$ncells_C4_other, p$ncells_C5_other))
+  )
+
+  for (sv in seeding_vars) {
+    if (sv$delay >= t_start && sv$delay <= t_end) {
+      # Turn on start flag
+      events[[ length(events) + 1 ]] <- data.frame(
+        var   = sv$flag,
+        time  = sv$delay,
+        value = 1,
+        method = "replace"
+      )
+      # Seed cancer cells
+      for (k in seq_along(sv$clones)) {
+        events[[ length(events) + 1 ]] <- data.frame(
+          var   = sv$clones[k],
+          time  = sv$delay,
+          value = sv$ncells[k],
+          method = "replace"
+        )
+      }
+    }
+  }
+
+  # ---------------------------------------------------------------------------
+  # 3. PRIMARY TUMOR SURGERY
+  # ---------------------------------------------------------------------------
+  if (p$do_surgery == 1 && p$surgery_time >= t_start && p$surgery_time <= t_end) {
+    for (ci in c("C1_T","C2_T","C3_T","C4_T","C5_T")) {
+      events[[ length(events) + 1 ]] <- data.frame(
+        var   = ci,
+        time  = p$surgery_time,
+        value = 0,
+        method = "replace"
+      )
+    }
+    events[[ length(events) + 1 ]] <- data.frame(
+      var   = "start",
+      time  = p$surgery_time,
+      value = 0,
+      method = "replace"
+    )
+  }
+
+  # Combine all events
+  if (length(events) == 0) return(NULL)
+  ev_df <- do.call(rbind, events)
+  ev_df <- ev_df[order(ev_df$time), ]
+  rownames(ev_df) <- NULL
+  return(ev_df)
+}
+
+# -----------------------------------------------------------------------------
+# event_function() for deSolve (alternative approach using eventfun)
+# This is called by ode() at each event time
+# -----------------------------------------------------------------------------
+make_event_function <- function(p, t_start = 0, t_end = 700) {
+
+  dose_nM    <- (p$dose_pembrolizumab / p$MW_pembrolizumab * 1e3) / p$V_C * 1e9
+  dose_times <- seq(t_start, t_end, by = p$dose_interval)
+
+  # Capture all needed values from p into the closure (avoids parms$ on atomic vector)
+  delay_Ln1    <- p$delay_Ln1
+  delay_Ln2    <- p$delay_Ln2
+  delay_other  <- p$delay_other
+  do_surgery   <- p$do_surgery
+  surgery_time <- p$surgery_time
+  seeding      <- p$seeding
+
+  ncells_C1_Ln1   <- p$ncells_C1_Ln1
+  ncells_C2_Ln1   <- p$ncells_C2_Ln1
+  ncells_C3_Ln1   <- p$ncells_C3_Ln1
+  ncells_C4_Ln1   <- p$ncells_C4_Ln1
+  ncells_C5_Ln1   <- p$ncells_C5_Ln1
+
+  ncells_C1_Ln2   <- p$ncells_C1_Ln2
+  ncells_C2_Ln2   <- p$ncells_C2_Ln2
+  ncells_C3_Ln2   <- p$ncells_C3_Ln2
+  ncells_C4_Ln2   <- p$ncells_C4_Ln2
+  ncells_C5_Ln2   <- p$ncells_C5_Ln2
+
+  ncells_C1_other <- p$ncells_C1_other
+  ncells_C2_other <- p$ncells_C2_other
+  ncells_C3_other <- p$ncells_C3_other
+  ncells_C4_other <- p$ncells_C4_other
+  ncells_C5_other <- p$ncells_C5_other
+
+  function(t, y, parms) {
+    # Pembrolizumab dose
+    if (any(abs(t - dose_times) < 0.01)) {
+      y["aPD1_C"] <- y["aPD1_C"] + dose_nM
+    }
+    # Metastatic seeding (Ln1)
+    if (abs(t - delay_Ln1) < 0.01 && seeding == 1) {
+      y["start_Ln1"] <- 1
+      y["C1_Ln1"] <- ncells_C1_Ln1
+      y["C2_Ln1"] <- ncells_C2_Ln1
+      y["C3_Ln1"] <- ncells_C3_Ln1
+      y["C4_Ln1"] <- ncells_C4_Ln1
+      y["C5_Ln1"] <- ncells_C5_Ln1
+    }
+    # Metastatic seeding (Ln2)
+    if (abs(t - delay_Ln2) < 0.01 && seeding == 1) {
+      y["start_Ln2"] <- 1
+      y["C1_Ln2"] <- ncells_C1_Ln2
+      y["C2_Ln2"] <- ncells_C2_Ln2
+      y["C3_Ln2"] <- ncells_C3_Ln2
+      y["C4_Ln2"] <- ncells_C4_Ln2
+      y["C5_Ln2"] <- ncells_C5_Ln2
+    }
+    # Metastatic seeding (other)
+    if (abs(t - delay_other) < 0.01 && seeding == 1) {
+      y["start_other"] <- 1
+      y["C1_oth"] <- ncells_C1_other
+      y["C2_oth"] <- ncells_C2_other
+      y["C3_oth"] <- ncells_C3_other
+      y["C4_oth"] <- ncells_C4_other
+      y["C5_oth"] <- ncells_C5_other
+    }
+    # Surgery
+    if (do_surgery == 1 && abs(t - surgery_time) < 0.01) {
+      y["C1_T"] <- 0; y["C2_T"] <- 0; y["C3_T"] <- 0
+      y["C4_T"] <- 0; y["C5_T"] <- 0; y["start"] <- 0
+    }
+    return(y)
+  }
+}
+
+# Helper: get all event times for ode() times argument
+get_event_times <- function(p, t_start = 0, t_end = 700) {
+  dose_times <- seq(t_start, t_end, by = p$dose_interval)
+  seed_times <- c(p$delay_Ln1, p$delay_Ln2, p$delay_other)
+  seed_times <- seed_times[seed_times >= t_start & seed_times <= t_end]
+  surg_time  <- if (p$do_surgery == 1) p$surgery_time else numeric(0)
+  all_times  <- sort(unique(c(dose_times, seed_times, surg_time)))
+  all_times[all_times >= t_start & all_times <= t_end]
+}