• Corpus ID: 220250505

Recommendations for machine learning validation in biology

  title={Recommendations for machine learning validation in biology},
  author={Ian Walsh and Dmytro Fishman and Dario Garc{\'i}a-Gasulla and Tiina Titma and Jennifer L. Harrow and Fotis Psomopoulos and Silvio C. E. Tosatto},
Modern biology frequently relies on machine learning to provide predictions and improve decision processes. There have been recent calls for more scrutiny on machine learning performance and possible limitations. Here we present a set of community-wide recommendations aiming to help establish standards of machine learning validation in biology. Adopting a structured methods description for machine learning based on DOME (data, optimization, model, evaluation) will allow both reviewers and… 

Figures and Tables from this paper

A Clinical Prognostic Model Based on Machine Learning from the Fondazione Italiana Linfomi (FIL) MCL0208 Phase III Trial

This is the first application of ML in a prospective clinical trial on MCL lymphoma and it is believed that ML would be of tremendous help in the development of a novel MCL prognostic score aimed at re-defining risk stratification.

Mitigation Strategies to Improve Reproducibility of Poverty Estimations From Remote Sensing Images Using Deep Learning

This study reports a review of the reproducibility of three DL experiments which analyze visual indicators from satellite and street imagery and proposes a checklist incorporating relevant FAIR principles to screen an experiment for its reproducecibility.



Correct machine learning on protein sequences: a peer-reviewing perspective

A set of guidelines to allow both peer reviewers and authors to avoid common machine learning pitfalls is espoused to help nonspecialists to appreciate the critical issues in machine learning.

Setting the standards for machine learning in biology

  • David T Jones
  • Computer Science
    Nature Reviews Molecular Cell Biology
  • 2019
The diverse applications of new ‘deep learning’ approaches with neural networks are now expanding into the field of biology but these applications to biological data require more scrutiny and caution to increase the standards of publishing and allow the AI revolution in biology to take off.

Machine learning applications in genetics and genomics

An overview of machine learning applications for the analysis of genome sequencing data sets, including the annotation of sequence elements and epigenetic, proteomic or metabolomic data is provided.

Validity of machine learning in biology and medicine increased through collaborations across fields of expertise

It is found that interdisciplinary collaborations increased the scientific validity of published research and suggested collaborations between computational and experimental scientists to generate more scientifically sound and impactful work integrating knowledge from both domains.

Applications of machine learning in drug discovery and development

The most useful techniques and how machine learning can promote data-driven decision making in drug discovery and development are discussed and major hurdles in the field are highlighted.

Assessing the accuracy of prediction algorithms for classification: an overview

We provide a unified overview of methods that currently are widely used to assess the accuracy of prediction algorithms, from raw percentages, quadratic error measures and other distances, and

Working toward precision medicine: Predicting phenotypes from exomes in the Critical Assessment of Genome Interpretation (CAGI) challenges

The range of techniques used for phenotype prediction as well as the methods used for assessing predictive models are discussed, and some of the difficulties associated with making predictions and evaluating them are outlined.

Machine Learning in Medicine.

  • R. Deo
  • Computer Science
  • 2015
What obstacles there may be to changing the practice of medicine through statistical learning approaches, and how these might be overcome are identified.

A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection

The results indicate that for real-word datasets similar to the authors', the best method to use for model selection is ten fold stratified cross validation even if computation power allows using more folds.