A systems microbiology framework for reproducible multi-dataset omics integration with application to long COVID.
7 blocks of code here have a permanent identifier (oscr:b:…), listed with the matches below: citable for good, with a page of its own. Select a match and press Cite to copy it. What a block identifier is.
The 8 matches
- [1] § Methods and materials › Nested validation and stability analysis ↔ Discovery/resultsDNP/dnp_run2.py, lines 1 to 34 · score 0.82 · layer gradient, activating features, feature selection, unselected, greedy, neural · block
oscr:b:2DVZF2JVK3181-D· Report a mismatch - [2] § Methods and materials › Nested validation and stability analysis ↔ Discovery/resultsGRACES/GRACESPortFinal.py, lines 17 to 33 · score 0.75 · gradient norms, dropout probability, graph, GRACES, epochs, weights · block
oscr:b:KGQKKMBQ6C5C9-I· Report a mismatch - [3] § Methods and materials › Nested validation and stability analysis ↔ src/features.py, lines 180 to 222 · score 0.70 · logistic regression, Random Forest, feature selection, trees, variables, Elastic · block
oscr:b:NBM1K2JDJX9YX-H· Report a mismatch - [4] § Methods and materials › Nested validation and stability analysis ↔ src/classifiersMulti.py, lines 71 to 113 · score 0.60 · logistic regression, Random Forest, trees, Elastic, LASSO, ensemble · block
oscr:b:F1Z7YX7ZA06Q8-L· Report a mismatch - [5] § Methods and materials › Nested validation and stability analysis ↔ Discovery/resultsDNP/dnp_run2.py, lines 59 to 94 · score 0.58 · ReLU, hidden layer, trained, MLP · block
oscr:b:H1J1BPW7Y8CNE-P· Report a mismatch - [6] § Methods and materials › Algorithmic framework for multi-dataset systems inference ↔ Discovery/resultsDNP/dnp_run2.py, lines 1 to 34 · score 0.55 · Deep Neural Pursuit, DNP · block
oscr:b:2DVZF2JVK3181-D· Report a mismatch - [7] § Results › GSE275334 (discovery dataset) ↔ Discovery/resultsRandomForestFI/selectRandomForestFeatureSelection.py, lines 16 to 27 · score 0.53 · Random Forest feature, feature selection, RF, classifier, discovery · block
oscr:b:KNE36EQVNM47Q-B· Report a mismatch - [8] § Results › GSE275334 (discovery dataset) ↔ Discovery/resultsBoruta/ROCMLP.py, lines 42 to 183 · score 0.50 · ROC curve, MLP classifier, fitting, MCC, discovery · block
oscr:b:4QGWKTWZJC1M7-5· Report a mismatch
Paper
Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC
The paper is loaded when this pane is shown.
The authors' code
Python · 514 lines · 18 KB · MIT · 3 matches
- #!/usr/bin/env python3
- # -*- coding: utf-8 -*-
- """
- Deep Neural Pursuit (DNP) — Single-file PyTorch implementation + main()
- Implements a DNP-style greedy feature selection for HDLSS data:
- • Greedy addition by input-layer gradient group-norms (L2) averaged across dropout samples
- • Keeps unselected input columns fixed at zero (W_C = 0) during subnetwork training
- • Adagrad optimizer; Xavier init for newly activated feature columns
- • Optional bootstrap stability (Jaccard)
- Inputs (CSV):
- - --data : shape (n_samples, n_features)
- - --features : shape (n_features,) one name per line
- - --labels : shape (n_samples,) binary labels {0,1}
- Primary outputs:
- - selected_genes_dnp.csv (ordered feature names)
- - dnp_report.txt (run summary)
- - stability_report.txt (if --stability-runs > 0)
- Additional outputs:
- - data_reduced.csv (X restricted to selected columns, same sample order)
- - features_reduced.csv (selected feature names, ordered)
- - indices_reduced.txt (0-based column indices, ordered)
- """
- import math
- import argparse
- import numpy as np
- import pandas as pd
- import torch
- import torch.nn as nn
- import torch.nn.functional as F
- #np.random.seed(42)
- # Optional: sklearn for AUC
- try:
- from sklearn.metrics import roc_auc_score
- SKLEARN_OK = True
- except Exception:
- SKLEARN_OK = False
- # -----------------------------
- # Utilities
- # -----------------------------
- def standardize_np(X):
- mu = X.mean(axis=0, keepdims=True)
- sd = X.std(axis=0, ddof=0, keepdims=True)
- sd[sd == 0] = 1.0
- return (X - mu) / sd, mu, sd
- def xavier_bound(fan_in, fan_out):
- return math.sqrt(6.0 / (fan_in + fan_out))
- # -----------------------------
- # MLP backbone (exposes first-layer weights)
- # -----------------------------
- class MLP(nn.Module):
- def __init__(self, input_dim, hidden_layers=(50, 30, 15), dropout_p=0.5):
- super().__init__()
- self.dropout_p = float(dropout_p)
- self.first = nn.Linear(input_dim, hidden_layers[0], bias=True)
- nn.init.zeros_(self.first.weight)
- nn.init.zeros_(self.first.bias)
- blocks = []
- prev = hidden_layers[0]
- for h in hidden_layers[1:]:
- blocks += [nn.Linear(prev, h), nn.ReLU(inplace=True), nn.Dropout(p=self.dropout_p)]
- prev = h
- self.hidden = nn.Sequential(*blocks)
- self.out = nn.Linear(prev, 1) # single logit
- nn.init.zeros_(self.out.weight)
- nn.init.zeros_(self.out.bias)
- def forward(self, x):
- z = self.first(x)
- z = F.relu(z)
- z = F.dropout(z, p=self.dropout_p, training=self.training)
- if len(self.hidden) > 0:
- z = self.hidden(z)
- logit = self.out(z)
- return logit.squeeze(1)
- @property
- def W1(self):
- # shape: [hidden1, input_dim]
- return self.first.weight
- # -----------------------------
- # DNP Classifier
- # -----------------------------
- class DNPClassifier:
- """
- PyTorch DNP:
- - Greedy selection via ||grad(W1[:, j])||_2 (averaged across dropout passes)
- - W_C=0 for unselected columns during training
- - Adagrad optimizer
- - Xavier init for newly activated column
- """
- def __init__(
- self,
- input_dim,
- hidden_layers=(50, 30, 15),
- k_features=25,
- dropout_p=0.5,
- n_dropout_samples=12,
- lr=1e-2,
- weight_decay=0.0,
- batch_size=None,
- max_epochs_per_iter=200,
- early_stopping_patience=20,
- device="auto",
- seed=42,
- ):
- self.input_dim = int(input_dim)
- self.hidden_layers = tuple(hidden_layers)
- self.k_features = int(k_features)
- self.dropout_p = float(dropout_p)
- self.n_dropout_samples = int(n_dropout_samples)
- self.lr = float(lr)
- self.weight_decay = float(weight_decay)
- self.batch_size = batch_size
- self.max_epochs_per_iter = int(max_epochs_per_iter)
- self.early_stopping_patience = int(early_stopping_patience)
- # Device resolution
- self.device = device
- if self.device in (None, "auto"):
- self.device = "cuda" if torch.cuda.is_available() else "cpu"
- self.seed = int(seed)
- torch.manual_seed(self.seed)
- if torch.cuda.is_available():
- torch.cuda.manual_seed_all(self.seed)
- self.model = MLP(self.input_dim, self.hidden_layers, dropout_p=self.dropout_p).to(self.device)
- self.selected_mask_ = np.zeros(self.input_dim, dtype=bool) # True if selected
- self.ranking_ = []
- self.scaler_ = None # (mu, sd)
- # -------- data helpers --------
- def _to_tensor(self, X, y=None):
- Xt = torch.as_tensor(X, dtype=torch.float32, device=self.device)
- if y is None:
- return Xt, None
- yt = torch.as_tensor(y, dtype=torch.float32, device=self.device)
- return Xt, yt
- def _loader(self, X, y):
- if self.batch_size is None or self.batch_size >= len(X):
- ds = torch.utils.data.TensorDataset(X, y)
- return torch.utils.data.DataLoader(ds, batch_size=len(X), shuffle=False)
- ds = torch.utils.data.TensorDataset(X, y)
- return torch.utils.data.DataLoader(ds, batch_size=self.batch_size, shuffle=True)
- # -------- W_C = 0 enforcement --------
- def _zero_candidate_cols(self):
- with torch.no_grad():
- W = self.model.W1 # [hidden1, input_dim]
- mask = torch.as_tensor(~self.selected_mask_, device=W.device) # True = candidate
- W[:, mask] = 0.0
- def _after_step_rezero(self):
- self._zero_candidate_cols()
- # -------- train one subnetwork --------
- def _train_subnetwork(self, Xtr, ytr, Xval, yval):
- self.model.train()
- opt = torch.optim.Adagrad(self.model.parameters(), lr=self.lr, weight_decay=self.weight_decay)
- bce = nn.BCEWithLogitsLoss()
- loader = self._loader(Xtr, ytr)
- best_val = float('inf')
- best_state = None
- wait = 0
- for _epoch in range(self.max_epochs_per_iter):
- for xb, yb in loader:
- opt.zero_grad(set_to_none=True)
- logits = self.model(xb)
- loss = bce(logits, yb)
- loss.backward()
- # Keep candidates frozen
- with torch.no_grad():
- W = self.model.W1
- mask_cand = torch.as_tensor(~self.selected_mask_, device=W.device)
- if W.grad is not None:
- W.grad[:, mask_cand] = 0.0
- opt.step()
- self._after_step_rezero()
- # Early stopping on validation loss
- self.model.eval()
- with torch.no_grad():
- vloss = bce(self.model(Xval), yval).item()
- self.model.train()
- if vloss < best_val - 1e-6:
- best_val = vloss
- best_state = {k: v.detach().cpu().clone() for k, v in self.model.state_dict().items()}
- wait = 0
- else:
- wait += 1
- if wait >= self.early_stopping_patience:
- break
- if best_state is not None:
- self.model.load_state_dict(best_state)
- self._after_step_rezero()
- # -------- candidate scoring: average dropout gradients --------
- def _score_candidates(self, X, y):
- self.model.train() # enable dropout
- bce = nn.BCEWithLogitsLoss()
- acc = torch.zeros_like(self.model.W1) # [hidden1, input_dim]
- # multiple gradient samplings with dropout noise
- for _ in range(self.n_dropout_samples):
- self._zero_candidate_cols()
- self.model.zero_grad(set_to_none=True)
- logits = self.model(X)
- loss = bce(logits, y)
- loss.backward()
- acc += self.model.W1.grad.clone()
- # average outside loop
- acc /= float(self.n_dropout_samples)
- # Column-wise L2 norms (feature scores)
- grad_cols = acc.norm(p=2, dim=0).detach().cpu().numpy() # [input_dim]
- # Guard against non-finite values
- grad_cols = np.where(np.isfinite(grad_cols), grad_cols, -np.inf)
- # Exclude already selected
- grad_cols[self.selected_mask_] = -np.inf
- # Tie-breaking jitter (seeded, tiny; keeps reproducibility but avoids argmax==0 ties)
- rng = np.random.RandomState(self.seed)
- eps = rng.normal(loc=0.0, scale=1e-12, size=grad_cols.shape)
- tie_safe = grad_cols + eps
- j = int(np.argmax(tie_safe))
- score = float(grad_cols[j])
- return j, score, grad_cols
- # -------- public API --------
- def fit(self, X, y, X_val=None, y_val=None, k=None, standardize=True):
- X = np.asarray(X, dtype=np.float32)
- y = np.asarray(y, dtype=np.float32).ravel()
- if standardize:
- X, mu, sd = standardize_np(X)
- self.scaler_ = (mu, sd)
- # Small internal validation split (20%) if none provided
- if X_val is None or y_val is None:
- n = len(X)
- m = max(1, int(0.2 * n))
- idx = np.arange(n)
- rng = np.random.RandomState(self.seed)
- rng.shuffle(idx)
- val_idx, tr_idx = idx[:m], idx[m:]
- Xtr, ytr = X[tr_idx], y[tr_idx]
- Xval, yval = X[val_idx], y[val_idx]
- else:
- Xtr, ytr = X, y
- Xval, yval = np.asarray(X_val, dtype=np.float32), np.asarray(y_val, dtype=np.float32)
- if standardize and self.scaler_ is not None:
- mu, sd = self.scaler_
- Xval = (Xval - mu) / sd
- Xtr_t, ytr_t = self._to_tensor(Xtr, ytr)
- Xval_t, yval_t = self._to_tensor(Xval, yval)
- self.ranking_ = []
- self.selected_mask_[:] = False
- target_k = int(k) if k is not None else self.k_features
- for _ in range(min(target_k, self.input_dim)):
- # 1) Train subnetwork on current selected set
- self._train_subnetwork(Xtr_t, ytr_t, Xval_t, yval_t)
- # 2) Score candidates via multi-dropout gradients
- j, _, _ = self._score_candidates(Xtr_t, ytr_t)
- # 3) Activate feature j
- self.selected_mask_[j] = True
- self.ranking_.append(j)
- # Xavier init for the new column
- fan_in = self.model.first.in_features
- fan_out = self.model.first.out_features
- bound = xavier_bound(fan_in, fan_out)
- with torch.no_grad():
- self.model.W1[:, j].uniform_(-bound, bound)
- self._zero_candidate_cols()
- return self
- @torch.no_grad()
- def predict_proba(self, X):
- X = np.asarray(X, dtype=np.float32)
- if self.scaler_ is not None:
- mu, sd = self.scaler_
- X = (X - mu) / sd
- Xt, _ = self._to_tensor(X, None)
- self.model.eval()
- logits = self.model(Xt)
- p = torch.sigmoid(logits).cpu().numpy()
- return np.vstack([1 - p, p]).T
- def selected_features_(self):
- return list(self.ranking_)
- # -----------------------------
- # CLI + main
- # -----------------------------
- def parse_args():
- p = argparse.ArgumentParser(description="Deep Neural Pursuit (DNP) — PyTorch")
- p.add_argument('--data', default='./data/data_0.csv', help='CSV: shape (n_samples, n_features)')
- p.add_argument('--features', default='./data/features_0.csv', help='CSV: feature names (n_features,)')
- p.add_argument('--labels', default='./data/labels.csv', help='CSV: labels (n_samples,)')
- p.add_argument('--k', type=int, default=7, help='Number of features to select')
- p.add_argument('--hidden', type=str, default='50,30,15', help='Hidden sizes, comma-separated')
- p.add_argument('--dropout', type=float, default=0.5, help='Dropout probability')
- p.add_argument('--dropout-samples', type=int, default=12, help='Dropout passes for gradient averaging')
- p.add_argument('--lr', type=float, default=1e-2, help='Learning rate')
- p.add_argument('--epochs', type=int, default=1000, help='Max epochs per greedy iteration')
- p.add_argument('--patience', type=int, default=20, help='Early stopping patience per iteration')
- p.add_argument('--device', type=str, default="auto", help='cpu/cuda/auto') # safer default
- p.add_argument('--seed', type=int, default=42, help='Random seed')
- p.add_argument('--standardize', action='store_true', help='Standardize features (z-score)')
- p.add_argument('--no-standardize', dest='standardize', action='store_false')
- p.set_defaults(standardize=True)
- p.add_argument('--stability-runs', type=int, default=0, help='Optional bootstrap runs for stability')
- p.add_argument('--output-prefix', type=str, default='selected_genes_dnp', help='Output prefix')
- return p.parse_args()
- def safe_auc(y_true, y_score):
- if not SKLEARN_OK:
- return None
- try:
- return float(roc_auc_score(y_true, y_score))
- except Exception:
- return None
- def jaccard(a, b):
- A, B = set(a), set(b)
- if not A and not B:
- return 1.0
- return len(A & B) / max(1, len(A | B)) # correct union
- def main():
- args = parse_args()
- # -----------------------------
- # 1. Load data
- # -----------------------------
- X = pd.read_csv(args.data, header=None).values.astype(np.float32)
- genes = pd.read_csv(args.features, header=None)[0].astype(str).tolist()
- y = pd.read_csv(args.labels, header=None)[0].values.astype(np.float32)
- assert X.shape[1] == len(genes), f"X has {X.shape[1]} cols, features file has {len(genes)} names"
- assert X.shape[0] == len(y), f"X has {X.shape[0]} rows, labels file has {len(y)} labels"
- # Convert to DataFrame for consistency with other scripts
- X_df = pd.DataFrame(X, columns=genes)
- hidden = tuple(int(x) for x in args.hidden.split(','))
- # -----------------------------
- # 2. Fit DNP
- # -----------------------------
- clf = DNPClassifier(
- input_dim=X.shape[1],
- hidden_layers=hidden,
- k_features=args.k,
- dropout_p=args.dropout,
- n_dropout_samples=args.dropout_samples,
- lr=args.lr,
- max_epochs_per_iter=args.epochs,
- early_stopping_patience=args.patience,
- device=args.device,
- seed=args.seed
- )
- clf.fit(X, y, standardize=args.standardize)
- # -----------------------------
- # 3. Extract selection
- # -----------------------------
- sel_idx = clf.selected_features_()
- sel_genes = [genes[j] for j in sel_idx]
- # -----------------------------
- # 4. Build feature ranking (FULL)
- # -----------------------------
- # You must have something like this in your classifier
- # e.g. gradient norms or importance scores
- if hasattr(clf, "feature_importances_"):
- scores = clf.feature_importances_
- elif hasattr(clf, "grad_norms_"):
- scores = clf.grad_norms_
- else:
- # fallback: zero importance except selected
- scores = np.zeros(len(genes), dtype=float)
- scores[sel_idx] = np.linspace(1.0, 0.5, len(sel_idx))
- importance_df = pd.DataFrame({
- "feature": genes,
- "importance": scores
- })
- importance_df = importance_df.sort_values(by="importance", ascending=False)
- # -----------------------------
- # 5. Reduce dataset
- # -----------------------------
- X_selected = X_df[sel_genes]
- print("Selected features:", len(sel_genes))
- print("Reduced shape:", X_selected.shape)
- # -----------------------------
- # 6. Save outputs (MATCH OTHERS)
- # -----------------------------
- X_selected.to_csv("X_dnp_selected.csv", index=False)
- importance_df.to_csv("dnp_feature_ranking.csv", index=False)
- pd.Series(sel_genes).to_csv(
- "dnp_selected_features.txt",
- index=False,
- header=False
- )
- print("\n✅ Files saved:")
- print(" - X_dnp_selected.csv")
- print(" - dnp_feature_ranking.csv")
- print(" - dnp_selected_features.txt")
- # -----------------------------
- # 7. Optional AUC (same as before)
- # -----------------------------
- proba_full = clf.predict_proba(X)[:, 1]
- auc_full = safe_auc(y, proba_full)
- if auc_full is not None:
- print(f"AUC (indicative, on full X): {auc_full:.4f}")
- # -----------------------------
- # 8. Optional stability (unchanged)
- # -----------------------------
- if args.stability_runs > 0:
- rng = np.random.RandomState(args.seed)
- selections = []
- for r in range(args.stability_runs):
- idx = rng.choice(np.arange(len(X)), size=len(X), replace=True)
- Xr, yr = X[idx], y[idx]
- dnp_r = DNPClassifier(
- input_dim=X.shape[1],
- hidden_layers=hidden,
- k_features=args.k,
- dropout_p=args.dropout,
- n_dropout_samples=args.dropout_samples,
- lr=args.lr,
- max_epochs_per_iter=args.epochs,
- early_stopping_patience=args.patience,
- device=args.device,
- seed=args.seed + r + 1
- )
- dnp_r.fit(Xr, yr, standardize=args.standardize)
- selections.append(dnp_r.selected_features_())
- jaccs = []
- for i in range(len(selections)):
- for j in range(i + 1, len(selections)):
- jaccs.append(jaccard(selections[i], selections[j]))
- stab_mean = float(np.mean(jaccs)) if jaccs else float('nan')
- with open('stability_report.txt', 'w', encoding='utf-8') as f:
- f.write("DNP Stability (Jaccard over selections)\n")
- f.write("=======================================\n")
- f.write(f"runs: {len(selections)}\n")
- f.write(f"mean_jaccard: {stab_mean:.4f}\n")
- print("Saved stability report: stability_report.txt")
- if __name__ == '__main__':
- main()
dnp_run2.py at commit 1123473, under MIT · at the source
Overview
- Informatics Institute, University of Amsterdam, Amsterdam, Netherlands
- Laboratory for the Design and Development of New Drugs and Biotechnological Innovation, Higher School of Medicine, National Polytechnic Institute (IPN), Mexico, Mexico
- Division of Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Faculty of Science, University of Utrecht, Utrecht, Netherlands
- Global Centre of Excellence Immunology, Danone Nutricia Research, Utrecht, Netherlands
Abstract
Integrative systems microbiology increasingly relies on algorithmic approaches capable of extracting biologically meaningful patterns from heterogeneous and often high dimensional, low-sample-size (HDLSS) biological datasets. A major obstacle in this setting is the instability of inferred molecular signatures across cohorts, tissues, and measurement platforms. Here, we address this problem by formulating molecular system inference as a multi-dataset integration task and by applying the Matthews Correlation Coefficient–Recursive Ensemble Feature Selection (MCC-REFS) algorithm to jointly analyze five independent transcriptomic datasets spanning peripheral blood mononuclear cells, whole blood, plasma, and post-mortem tissues. We compared MCC-REFS with three commonly used feature-selection strategies, GRACES, SelectKBest, and Deep Neural Pursuit (DNP), in order to evaluate robustness, convergence, and cross-context reproducibility. MCC-REFS consistently converged on a compact seven-gene system (PPP2CB, SOCS3, ARG1, IL6R, ECHS1, FZD2, TRGV3/
Reproduced under the paper's license (CC BY), from the paper cited above.
Repository
Its files are read in the Code ↔ Paper reader above, with 8 matches between paragraphs and lines of code.
steppenwolf0/longCOVIDMCCREFS
11234733188788c3cc65dce8ef0c5363760cb855, 27 July 2026Availability: 1 check, the latest on 26 September 2026: the link answers
- 26 September 2026: the link answers
35 files
- Discovery/
resultsBoruta/ , Python, 187 lines, 1 matchROCMLP.py - Discovery/
resultsBoruta/ , Python, 75 linesselectBoruta.py - Discovery/
resultsDNP/ , Python, 187 linesROCMLP.py - Discovery/
resultsDNP/ , Python, 514 lines, 3 matchesdnp_run2.py - Discovery/
resultsEN/ , Python, 187 linesROCMLP.py - Discovery/
resultsEN/ , Python, 88 linesselectEN.py - Discovery/
resultsGRACES/ , Python, 178 linesGRACES.py - Discovery/
resultsGRACES/ , Python, 387 lines, 1 matchGRACESPortFinal.py - Discovery/
resultsGRACES/ , Python, 187 linesROCMLP.py - Discovery/
resultsKbest/ , Python, 187 linesROCMLP.py - Discovery/
resultsKbest/ , Python, 97 linesselectKBest.py - Discovery/
resultsLASSO/ , Python, 187 linesROCMLP.py - Discovery/
resultsLASSO/ , Python, 84 linesselectLASSO.py - Discovery/
resultsMCCREFS2/ , Python, 187 linesROCMLP.py - Discovery/
resultsRFECV/ , Python, 187 linesROCMLP.py - Discovery/
resultsRFECV/ , Python, 89 linesselectRFECV.py - Discovery/
resultsRandomForestFI/ , Python, 187 linesROCMLP.py - Discovery/
resultsRandomForestFI/ , Python, 74 lines, 1 matchselectRandomForestFeatur eSelection.py - Discovery/
resultsSHAP/ , Python, 187 linesROCMLP.py - Discovery/
resultsSHAP/ , Python, 93 linesselectShap.py - E-ENAD-46/
ROCMLP.py , Python, 186 lines - GSE157103/
ROCMLP.py , Python, 186 lines - GSE226260/
Dataset1/ , Python, 25 linessubsetDataset.py - GSE226260/
Dataset2/ , Python, 25 linessubsetDataset.py - GSE226260/
ROCMLP.py , Python, 186 lines - GSE226260/
subsetDataset.py , Python, 25 lines - GSE270045/
ROCMLP.py , Python, 186 lines - src/
aBioInf100.py , Python, 51 lines - src/
classifiersMulti.py , Python, 232 lines, 1 match - src/
features.py , Python, 357 lines, 1 match - src/
reduceData.py , Python, 49 lines - src/
sumFig.py , Python, 91 lines - src/
summaryMulti.py , Python, 123 lines - LICENSE, License, 21 lines
- README.md, Text, 26 lines
The paper's code and data availability statement is in the Data section.
Tracing map
Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.
What the map holds:
- 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
- 33 scripts, each with its path and the digest of its content;
- 8 matches between paragraphs of the paper and lines of the code (method lexical-v1);
- neither the text of the paper nor the code itself.
Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.
Data
Datasets cited
- geo:GSE157103, at NCBI GEO; found in “Data availability statement”
Data availability statement
The datasets analyzed in this study are publicly available from established repositories. Transcriptomic datasets GSE275334 (https://
Reproduced under the paper's license (CC BY), from the paper cited above.
Versions
The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.
Version 1, 27 September 2026: the first record
Recorded: type, language, journal, volume, pages, dates, 5 authors, 6 keywords, 64 references.
Cite
This paper
Varga, B., Martinez-Archundia, M., Willemsen, L. E. M., Garssen, J., & Lopez-Rincon, A. (2026). A systems microbiology framework for reproducible multi-dataset omics integration with application to long COVID. Frontiers in systems biology, 6, 1873899. https://
BibTeX
@article{varga2026system
author = {Varga, Brigitta and Martinez-Archundia, Marlet and Willemsen, Linette E M and Garssen, Johan and Lopez-Rincon, Alejandro},
title = {{A systems microbiology framework for reproducible multi-dataset omics integration with application to long COVID}},
journal = {Frontiers in systems biology},
year = {2026},
month = sep,
volume = {6},
pages = {1873899},
publisher = {Frontiers Media},
issn = {2674-0702},
doi = {10.3389/
url = {https://
pmid = {42755588},
pmcid = {PMC13581729}
}
RIS
TY - JOUR
AU - Varga, Brigitta
AU - Martinez-Archundia, Marlet
AU - Willemsen, Linette E M
AU - Garssen, Johan
AU - Lopez-Rincon, Alejandro
TI - A systems microbiology framework for reproducible multi-dataset omics integration with application to long COVID
T2 - Frontiers in systems biology
J2 - Front Syst Biol
PY - 2026
DA - 2026/
VL - 6
SP - 1873899
SN - 2674-0702
PB - Frontiers Media
DO - 10.3389/
UR - https://
LA - en
ER -
CSL-JSON
{
"id": "10.3389/
"type": "article-journal",
"title": "A systems microbiology framework for reproducible multi-dataset omics integration with application to long COVID",
"container-title": "Frontiers in systems biology",
"author": [
{
"family": "Varga",
"given": "Brigitta"
},
{
"family": "Martinez-Archundia",
"given": "Marlet"
},
{
"family": "Willemsen",
"given": "Linette E M"
},
{
"family": "Garssen",
"given": "Johan"
},
{
"family": "Lopez-Rincon",
"given": "Alejandro"
}
],
"container-title-short":
"volume": "6",
"page": "1873899",
"DOI": "10.3389/
"PMID": "42755588",
"PMCID": "PMC13581729",
"ISSN": "2674-0702",
"publisher": "Frontiers Media",
"URL": "https://
"language": "en",
"issued": {
"date-parts": [
[
2026,
9,
3
]
]
}
}
The tracing map gets a citation of its own once an author has validated it and it has a DOI.
Similar papers
The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.
- [1] doi:10.1093/bib/bbae250 [code]
- Benchmarking mapping algorithms for cell-type annotating in mouse brain by integrating single-nucleus RNA-seq and Stereo-seq data.Journal: Briefings in bioinformaticsIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools, methods / tools, genetics / omics
- [2] doi:10.1038/s41467-025-57027-x [code]
- Topological identification and interpretation for single-cell epigenetic regulation elucidation in multi-tasks using scAGDE.Journal: Nature communicationsIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools, genetics / omics
- [3] doi:10.1016/j.xgen.2024.100739 [code]
- CSI-GEP: A GPU-based unsupervised machine learning approach for recovering gene expression programs in atlas-scale single-cell RNA-seq data.Journal: Cell genomicsIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools, genetics / omics
- [4] doi:10.1016/j.xgen.2024.100581 [code]
- SIMS: A deep-learning label transfer tool for single-cell RNA sequencing analysis.Journal: Cell genomicsIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools, genetics / omics
- [5] doi:10.1016/j.isci.2024.109027 [code]
- A deep learning adversarial autoencoder with dynamic batching displays high performance in denoising and ordering scRNA-seq data.Journal: iScienceIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools, genetics / omics
- [6] doi:10.1523/eneuro.0362-25.2026 [code]
- Similarities between &
lt;i& gt;Ciona& lt;/ i& gt; Dorsal Motor Ganglion and Vertebrate Cerebellum: Did a Chordate Ancestor Already Show D/ V Subdivision within a Hindbrain Precursor? Journal: eNeuroIn common: PyTorch Geometric, SHAP, statsmodels, 6 other tools - [7] doi:10.3389/fphar.2023.1291246 [code]
- Development and validation of PAMPA-BBB QSAR model to predict brain penetration potential of novel drug candidates.Journal: Frontiers in pharmacologyIn common: PyTorch Geometric, SHAP, PyTorch, 5 other tools, methods / tools
- [8] doi:10.1016/j.csbj.2024.11.013 [code]
- Topological embedding and directional feature importance in ensemble classifiers for multi-class classification.Journal: Computational and structural biotechnology journalIn common: SHAP, statsmodels, PyTorch, 5 other tools, 1 reference
- [9] doi:10.1186/s12864-025-11946-7 [code]
- GNODEVAE: a graph-based ODE-VAE enhances clustering for single-cell data.Journal: BMC genomicsIn common: PyTorch Geometric, statsmodels, PyTorch, 5 other tools, methods / tools, genetics / omics
- [10] doi:10.1038/s41467-023-44367-9 [code]
- MENDER: fast and scalable tissue structure identification in spatial omics data.Journal: Nature communicationsIn common: PyTorch Geometric, statsmodels, PyTorch, 5 other tools, methods / tools, genetics / omics
Contribute
The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.
Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.
Claim this paper
Correct its record
Say what each link of this record is, remove the ones that are not the paper's, add the ones that are missing. The correction becomes a new version of the record, in its Versions section.
Validate its tracing map
You validate the map as this page shows it: 1 repository of the authors' code, each at its verified commit and with its license, 33 scripts, and 8 matches between paragraphs and code (see the Code and Map sections). It then receives a DOI on Zenodo, with you (your ORCID iD) and OSCR as its creators; the code itself is not deposited.
The map's fingerprint: sha256:3123f923b3383640…
Add the badge to the code's README
If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.
Markdown
<!-- oscr-badge:start -->
[](https://openscicode.org/badge/doi/10.3389/fsysb.2026.1873899)
<!-- oscr-badge:end -->reStructuredText
.. oscr-badge:start
.. image:: https://openscicode.org/badge/doi/10.3389/fsysb.2026.1873899.svg
:target: https://openscicode.org/badge/doi/10.3389/fsysb.2026.1873899
:alt: OSCR
.. oscr-badge:endHTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.3389/fsysb.2026.1873899"><img src="https://openscicode.org/badge/doi/10.3389/fsysb.2026.1873899.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".
If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge page.
Request its removal
To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).
Discussion, reproductions, activity
Discussion: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.
Report a mismatch between the code and the paper · Report a code error (signed in)
Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.
Report a reproduction failure (signed in)
Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.
Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.
