{"id":46893658,"url":"https://github.com/bibymaths/phoscrosstalk","last_synced_at":"2026-03-10T23:21:58.058Z","repository":{"id":326458322,"uuid":"1105626366","full_name":"bibymaths/phoscrosstalk","owner":"bibymaths","description":"ODE-based phosphorylation network model using PTMcode2/pynetphorest crosstalk predictions to build global/local coupling matrices. Fits phosphosite time-course data with DE optimization and reconstructs full protein–site activation dynamics.","archived":false,"fork":false,"pushed_at":"2026-02-04T23:26:25.000Z","size":447,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-02-05T10:52:32.427Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"bsd-3-clause","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/bibymaths.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-11-27T23:05:09.000Z","updated_at":"2025-12-07T01:15:14.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/bibymaths/phoscrosstalk","commit_stats":null,"previous_names":["bibymaths/phoscrosstalk"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/bibymaths/phoscrosstalk","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bibymaths%2Fphoscrosstalk","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bibymaths%2Fphoscrosstalk/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bibymaths%2Fphoscrosstalk/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bibymaths%2Fphoscrosstalk/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bibymaths","download_url":"https://codeload.github.com/bibymaths/phoscrosstalk/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bibymaths%2Fphoscrosstalk/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30361567,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-10T21:41:54.280Z","status":"ssl_error","status_checked_at":"2026-03-10T21:40:59.357Z","response_time":106,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.5:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-03-10T23:21:57.575Z","updated_at":"2026-03-10T23:21:58.052Z","avatar_url":"https://github.com/bibymaths.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cimg src=\"images/phoscrosstalk_logo.svg\" width=\"450\"\u003e\n\n\u003cp align=\"left\"\u003e\n\n\u003cimg src=\"[https://img.shields.io/badge/python-3.10%2B-3776ab?logo=python\u0026logoColor=white](https://img.shields.io/badge/python-3.10%2B-3776ab?logo=python\u0026logoColor=white)\"\u003e\n\n\u003cimg src=\"[https://img.shields.io/badge/License-BSD--3--Clause-2ea043](https://img.shields.io/badge/License-BSD--3--Clause-2ea043)\"\u003e\n\n\u003cimg src=\"[https://img.shields.io/badge/ODE--Model-Systems%20Biology-blueviolet](https://img.shields.io/badge/ODE--Model-Systems%20Biology-blueviolet)\"\u003e\n\n\u003cimg src=\"[https://img.shields.io/badge/Optimization-NSGA3%20%7C%20UNSGA3-ff69b4](https://www.google.com/search?q=https://img.shields.io/badge/Optimization-NSGA3%2520%257C%2520UNSGA3-ff69b4)\"\u003e\n\n\u003c/p\u003e\n\n# **PhosCrosstalk**\n\n**Global phospho-network ODE modeling with PTM-based crosstalk integration and multi-objective evolutionary optimization\n**\n\nPhosCrosstalk is a systems-level phosphorylation modeling framework that integrates **PTMcode2-derived inter/intra\ncrosstalk**, **KEA3 kinase-substrate networks**, and **experimental phosphosite time-series** into a single unified *\n*global ODE model**.\nIt reconstructs protein activation, kinase activity, and phosphosite kinetics across an entire network, using **parallel\nMulti-Objective Evolutionary Algorithms (MOEAs)** via `pymoo` to fit large parameter sets efficiently and robustly.\n\nPhosCrosstalk provides a full end-to-end pipeline:\n\n* **Automated Data Curation**: Downloads and standardizes KEA, PhosphoSitePlus, and PTMcode2 data.\n* **Network Construction**: Builds multiplex kinase graphs and functional crosstalk matrices.\n* **Global Optimization**: Fits kinetic parameters using advanced MOEAs (NSGA-II, UNSGA-III).\n* **Simulation \u0026 Analysis**: Runs steady-state convergence, in-silico knockouts, and global sensitivity analysis (\n  Sobol).\n* **Interactive Visualization**: Includes a Streamlit dashboard for exploring trajectories and dynamic network\n  animations.\n\n---\n\n---\n\n# **Key Features**\n\n### **1. Global ODE Phospho-Network Model**\n\nThe model captures the dynamics of three interconnected biological layers:\n\n* **Protein Activation (`S`)**: Fraction of active protein.\n* **Kinase Activity (`K_dyn`)**: Dynamic activity of kinases, regulated by upstream inputs and network topology.\n* **Phosphosite Occupancy (`p`)**: Fractional phosphorylation of specific residues.\n\nThe coupled ODE system integrates:\n\n* **Kinase-Substrate Interactions**: Directional phosphorylation driven by kinase activity (`K_dyn`).\n* **Global Crosstalk**: Functional coupling from PTMcode2 inter/intra-protein associations (`β_g * Cg`).\n* **Local Proximity**: Sequence-based influence between nearby residues (`β_l * Cl`).\n* **Mechanistic Flexibility**: Supports **Distributive**, **Sequential**, and **Random/Cooperative** kinetic mechanisms.\n\n---\n\n### **2. Automated Data Curation Pipeline**\n\nA built-in curator module (`data_curator.py`) handles the heavy lifting of data acquisition:\n\n* **Downloads** raw datasets from Harmonizome (KEA, PhosphoSitePlus).\n* **Processes** PTMcode2 files into optimized SQLite databases.\n* **Constructs** a unified Kinase-Kinase interaction graph (NetworkX/Pickle).\n* **Maps** Kinase-Substrate relationships into fast lookup indices.\n\n---\n\n### **3. Multi-Objective Evolutionary Optimization**\n\nPhosCrosstalk uses `pymoo` to solve a multi-objective problem, simultaneously minimizing:\n\n1. **Phosphosite Error**: Difference between simulated and observed phosphorylation profiles.\n2. **Protein Abundance Error**: Difference between simulated and observed protein levels.\n3. **Model Complexity**: Regularization terms (L2 and Laplacian network smoothing).\n\nStrategies include **NSGA-II** (diversity-focused) and **UNSGA-III** (convergence-focused), run in parallel to find\nrobust Pareto-optimal solutions.\n\n---\n\n### **4. Advanced Post-Optimization Analysis**\n\nBeyond simple fitting, the framework offers deep analytical tools:\n\n* **Steady-State Analysis**: Simulates long-term convergence ().\n* **In-Silico Knockouts**: Systematically perturbs kinases, proteins, or sites to predict network-wide impacts (Fold\n  Change analysis).\n* **Global Sensitivity Analysis (GSA)**: Computes Sobol indices to identify high-impact parameters.\n* **Fréchet Distance Selection**: Selects the biologically \"best\" trajectory from the Pareto front.\n\n---\n\n### **5. Interactive Dashboard**\n\nA comprehensive **Streamlit** app allows you to:\n\n* Visualize fitted trajectories vs. experimental data.\n* Explore sensitivity rankings and parameter distributions.\n* Run real-time knockout simulations.\n* **Animate** the flow of kinase activity through the network over time.\n\n---\n\n---\n\n# **Repository Structure**\n\n```\nphoscrosstalk/\n│\n├── __init__.py\n├── main.py                     # Entry point for modeling \u0026 optimization\n├── data_curator.py             # Pipeline for downloading \u0026 processing raw data\n├── core_mechanisms.py          # Numba-accelerated ODE kernels\n├── optimization.py             # Pymoo Problem definitions \u0026 objectives\n├── simulation.py               # Scipy odeint wrappers\n├── analysis.py                 # Post-processing \u0026 static plotting\n├── sensitivity.py              # SALib Global Sensitivity Analysis\n├── knockouts.py                # Systematic in-silico perturbation screens\n├── app.py                      # Interactive Streamlit Dashboard\n│\n└── README.md\n\n```\n\n---\n\n# **Installation**\n\nPhosCrosstalk requires Python ≥ 3.10.\n\n```bash\ngit clone https://github.com/\u003cyourname\u003e/phoscrosstalk.git\ncd phoscrosstalk\n\npython -m venv venv\nsource venv/bin/activate\n\npip install -r requirements.txt\n\n```\n\n**Key Dependencies:** `numpy`, `scipy`, `pandas`, `numba`, `pymoo`, `networkx`, `salib`, `streamlit`, `rich`.\n\n---\n\n# **Data Curation**\n\nBefore modeling, you must curate the biological prior knowledge.\n\n### **1. Download Manual Files**\n\nDownload the **PTMcode2** within/between files from\nthe [PTMcode website](https://www.google.com/search?q=https://ptmcode.embl.de/downloads.cgi) and place them in a\nfolder (e.g., `data/ptmcode2/`).\n\n### **2. Run the Curator**\n\nThis command downloads KEA/PSP data automatically and processes your PTMcode files:\n\n```bash\npython3 -m phoscrosstalk.data_curator \\\n  --all \\\n  --ptmcode data/ptmcode2/within.gz data/ptmcode2/between.gz\n\n```\n\n*Outputs are saved to `data_curated/processed/`.*\n\n---\n\n# **Usage**\n\n### **Run the Modeling Pipeline**\n\nExecute the main optimization routine using your time-series data and the curated artifacts:\n\n```bash\nphoscrosstalk \\\n  --data data_timeseries/filtered_input1.csv \\\n  --ptm-intra data_curated/processed/ptm_intra.db \\\n  --ptm-inter data_curated/processed/ptm_inter.db \\\n  --kea-ks-table data_curated/processed/ks_psite_table.tsv \\\n  --unified-graph-pkl data_curated/processed/unified_kinase_graph.gpickle \\\n  --outdir results/experiment_01 \\\n  --cores 16 \\\n  --mechanism rand \\\n  --gen 300 \\\n  --run-steadystate \\\n  --run-knockouts \\\n  --run-sensitivity\n\n```\n\n### **Run the Dashboard**\n\nExplore the results interactively:\n\n```bash\nstreamlit run phoscrosstalk/app.py\n\n```\n\n*(Point the sidebar to your `results/experiment_01` directory)*\n\n---\n\n# **Output Files**\n\nThe pipeline generates a rich set of results in the output directory:\n\n* **`fit_timeseries.tsv`**: Long-format table of Observed vs. Simulated values for all sites.\n* **`fitted_params.npz`**: Complete archive of optimized parameters and model state.\n* **`pareto_front_with_J.tsv`**: Objective values for all solutions on the Pareto front.\n* **`knockouts/`**: Tables and heatmaps of Fold Changes for every in-silico knockout.\n* **`sensitivity/`**: Sobol indices (`sobol_indices_labeled.tsv`) and perturbation trajectories.\n* **`equations/`**: Automatically generated LaTeX report of the specific ODE system fitted.\n\n---\n\n# **Why PhosCrosstalk Exists**\n\nPhosphorylation is not isolated. Sites influence each other across:\n\n* protein domains\n* protein complexes\n* signaling cascades\n* PTM interaction networks\n\nMost modeling approaches treat sites independently or only use kinase–substrate data.\nPhosCrosstalk closes the gap: it integrates **global PTM relationships**, **local sequence context**, and **experimental\ntime-series**, giving a mechanistic, quantitative reconstruction of network-level phosphorylation dynamics.\n\nThis creates a bridge between:\n\n✔ dynamic ODE modeling\n✔ phosphoproteomics\n✔ PTM curation databases\n✔ machine-learning residue prediction tools\n\n---\n\n# **Citation**\n\n1. **Casado, P.,** Rodriguez-Prados, J.-C., Cosulich, S. C., Guichard, S., Vanhaesebroeck, B., \u0026 Cutillas, P. R. (2013).\n   Kinase-Substrate Enrichment Analysis provides insights into the heterogeneity of signaling pathway activation in\n   leukemia cells. *Science Signaling*, *6*(264),\n   rs6. [https://doi.org/10.1126/scisignal.2003573](https://doi.org/10.1126/scisignal.2003573)\n2. **Hornbeck, P. V.,** Zhang, B., Murray, B., Kornhauser, J. M., Latham, V., \u0026 Skrzypek, E. (2015). PhosphoSitePlus,\n   2014: mutations, PTMs and recalibrations. *Nucleic Acids Research*, *43*(D1),\n   D512–D520. [https://doi.org/10.1093/nar/gku1267](https://doi.org/10.1093/nar/gku1267)\n3. **Horn, H.,** Schoof, E., Kim, J., Robin, X., Miller, M. L., Diella, F., Palma, A., Cesareni, G., Jensen, L. J., \u0026\n   Linding, R. (2014). KinomeXplorer: an integrated platform for kinome biology studies. *Nature Methods*, *11*(6),\n   603–604. [https://doi.org/10.1038/nmeth.2968](https://doi.org/10.1038/nmeth.2968)\n4. **Minguez, P.,** Letunic, I., Parca, L., \u0026 Bork, P. (2013). PTMcode: a database of known and predicted functional\n   associations between post-translational modifications in proteins. *Nucleic Acids Research*, *41*(D1),\n   D306–D311. [https://doi.org/10.1093/nar/gks1230](https://doi.org/10.1093/nar/gks1230)\n5. **Linding, R.,** Jensen, L. J., Pasculescu, A., Olhovsky, M., Colwill, K., Bork, P., Yaffe, M. B., \u0026 Pawson, T. (\n   2008). NetworKIN: a resource for exploring cellular phosphorylation networks. *Nucleic Acids Research*, *36*(Database\n   issue),\n   D695–D699. [https://doi.org/10.1093/nar/gkm902](https://www.google.com/search?q=https://doi.org/10.1093/nar/gkm902)","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbibymaths%2Fphoscrosstalk","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbibymaths%2Fphoscrosstalk","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbibymaths%2Fphoscrosstalk/lists"}