---
name: machine-learning-model-training-sklearn
description: Use when your metabolomics analysis pipeline requires CCS value prediction for ion-mobility mass spectrometry data, you have access to a curated training set of known metabolites with experimentally validated CCS values, and you plan to predict CCS values on target datasets containing 10,000+.
license: CC-BY-4.0
metadata:
  edam_operation: http://edamontology.org/operation_3501
  edam_topics:
  - http://edamontology.org/topic_3520
  - http://edamontology.org/topic_0092
  tools:
  - Sklearn
  - Scikit-learn
  - Jupyter Lab
  - Python
  - CCSP 2.0
  techniques:
  - ion-mobility-MS
derived_from:
- doi: 10.1021/acs.analchem.2c03491
  title: CCS Predictor 2.0
- doi: 10.1101/2022.08.09.503345
  title: ''
evidence_spans:
- The current version of CCSP 2.0 requires Sklearn V1.0.2 or later
claims: []
provenance:
  collection: https://w3id.org/holobiomicslab/asb-skill/collection/metabolomics/v2
  assembled_by: scripts/collect_metabolomics_collection.py
  sources:
  - build: coll_ccs_predictor_2_0_cq
    doi: 10.1021/acs.analchem.2c03491
    title: CCS Predictor 2.0
  dedup_kept_from: coll_ccs_predictor_2_0_cq
schema_version: 0.2.0
attribution:
  generator: AgenticScienceBuilder
  original_doi: 10.1021/acs.analchem.2c03491
  all_source_dois:
  - 10.1021/acs.analchem.2c03491
  - 10.1101/2022.08.09.503345
  zenodo_doi: 10.5281/zenodo.20794027
  curators: []
  promoter: Louis-Félix Nothias
  sponsor: CNRS & Université Côte d'Azur
---

# machine-learning-model-training-sklearn

## Summary

Train a Scikit-learn collision cross section (CCS) prediction model on user-curated ion-mobility metabolomics training sets to enable accurate CCS value prediction and filtering of false positive identifications. This skill targets practitioners working with large-scale metabolomics datasets who need to integrate machine learning into ion-mobility workflows.

## When to use

Your metabolomics analysis pipeline requires CCS value prediction for ion-mobility mass spectrometry data, you have access to a curated training set of known metabolites with experimentally validated CCS values, and you plan to predict CCS values on target datasets containing 10,000+ molecules or integrate predictions into a larger local analysis workflow.

## When NOT to use

- Your training set is smaller than ~100 metabolites with reliable CCS annotations; insufficient data will produce unreliable model generalizations.
- You are performing small-scale predictions (< 10,000 molecules) and do not need local execution; Google Colaboratory is more suitable and requires no Python installation.
- Your metabolomics data is not ion-mobility mass spectrometry; CCS prediction is specific to drift-time or traveling-wave ion-mobility platforms.

## Inputs

- Curated training set: tabular data (CSV/TSV) with metabolite identifiers, molecular features, and experimentally measured CCS values
- Target metabolomics dataset: ion-mobility mass spectrometry data for which CCS values are to be predicted
- Feature set specification: selection of input molecular descriptors or structural properties for model training

## Outputs

- Trained Sklearn machine learning model object (serialized in the notebook environment)
- Predicted CCS values for target metabolites
- Filtered results table with false positive annotations removed
- Validation comparison: predicted vs. reference CCS values on benchmark datasets

## How to apply

Load the CCSP 2.0 Jupyter Lab notebook template and configure Scikit-learn (v1.0.2 or later) in your local Python environment. Upload your training dataset via the Tk graphical interface and select your input feature set (molecular descriptors or structural properties relevant to ion-mobility behavior). Train the Sklearn machine learning model on the user-provided training set using the notebook's embedded algorithms. Execute predictions on your target metabolomics dataset and filter the output results to remove annotations flagged as false positives. Validate that model predictions on reference datasets match expected CCS outputs to confirm correct training and parameter configuration.

## Related tools

- **Scikit-learn** (Machine learning library for training and executing the CCS prediction model; v1.0.2 or later required) — https://scikit-learn.org
- **Jupyter Lab** (Interactive notebook environment for running the CCSP 2.0 notebook with embedded Tk interface for data upload and parameter configuration) — https://jupyter.org
- **Python** (Programming language in which CCSP 2.0 is written and executed)
- **CCSP 2.0** (Complete Jupyter notebook package integrating Sklearn model training, data I/O via Tk interface, and false positive filtering for CCS prediction) — https://github.com/facundof2016/CCSP2.0

## Evaluation signals

- Trained model converges without errors on the curated training set; Sklearn training loss decreases monotonically.
- Predicted CCS values on reference benchmark molecules fall within ±2–5% error margin of experimentally measured values (standard for ion-mobility prediction).
- Output false positive filter flag correctly identifies and removes metabolite identifications with low CCS prediction confidence.
- Notebook execution completes without package version conflicts; Sklearn version ≥ 1.0.2 is loaded successfully.
- Predicted CCS values for target dataset are numeric, non-null, and biologically plausible (within observed CCS range of reference training set).

## Limitations

- Model prediction accuracy depends critically on the quality and representativeness of the curated training set; biased or small training sets will degrade generalization.
- Sklearn-based predictions are not inherently interpretable; the notebook does not provide feature importance or mechanistic insight into which molecular properties drive CCS predictions.
- Google Colaboratory execution is limited to 12 hours continuous operation and ~20 minutes idle timeout; only use for small-scale predictions (< 10,000 molecules).
- Large training or target datasets must be processed locally via Jupyter Lab; cloud-hosted options incur timeout and connectivity penalties.
- The Tk interface is platform-dependent; compatibility may vary across Windows, macOS, and Linux systems.

## Evidence

- [readme] Collision Cross Section Predictor 2.0 is an open source Python notebook intended to help ion-mobility scientists predict collision cross sections with user-curated training sets: "Collision Cross Section Predictor 2.0 is an open source Python notebook intended to help ion-mobility scientists predict collision cross sections with user-curated training sets"
- [readme] CCSP 2.0 is written in Python and packaged into two notebook forms: (1) a Google Colaboratory Jupyter notebook that is well suited for beginners, and (2) a Jupyter Lab compatible notebook with a Tk interface for users more familiar with Python.: "CCSP 2.0 is written in Python and packaged into two notebook forms: (1) a Google Colaboratory Jupyter notebook that is well suited for beginners, and (2) a Jupyter Lab compatible notebook with a Tk"
- [readme] The current version of CCSP 2.0 requires Sklearn V1.0.2 or later: "The current version of CCSP 2.0 requires Sklearn V1.0.2 or later"
- [readme] If you plan to predict CCS values using large training or target sets or if you plan to integrate the code into a larger analysis workflow, the locally hosted Jupyter Lab version is recommended.: "If you plan to predict CCS values using large training or target sets or if you plan to integrate the code into a larger analysis workflow, the locally hosted Jupyter Lab version is recommended"
- [other] Integrate the Tk graphical interface components to allow users to upload training data, configure model parameters, and select input feature sets.: "Integrate the Tk graphical interface components to allow users to upload training data, configure model parameters, and select input feature sets"
- [other] Implement the collision cross section prediction model using Sklearn's machine learning algorithms on the user-provided training set.: "Implement the collision cross section prediction model using Sklearn's machine learning algorithms on the user-provided training set"
- [other] Execute predictions on target metabolomics data and output filtered results with false positive annotations.: "Execute predictions on target metabolomics data and output filtered results with false positive annotations"
- [readme] If you plan to use CCSP 2.0 only to make small scale predictions (<10,000 molecules) and export your results to your local computer, the Google Colab option is recommended.: "If you plan to use CCSP 2.0 only to make small scale predictions (<10,000 molecules) and export your results to your local computer, the Google Colab option is recommended"