Active machine learning-driven experimentation to determine compound effects on protein patterns

  1. Armaghan W Naik
  2. Joshua D Kangas
  3. Devin P Sullivan
  4. Robert F Murphy  Is a corresponding author
  1. Carnegie Mellon University, United States

Abstract

High throughput screening determines the effects of many conditions on a given biological target. Currently, to estimate the effects of those conditions on other targets requires either strong modeling assumptions (e.g. similarities among targets) or separate screens. Ideally, data-driven experimentation could be used to learn accurate models for many conditions and targets without doing all possible experiments. We have previously described an active machine learning algorithm that can iteratively choose small sets of experiments to learn models of multiple effects. We now show that, with no prior knowledge and with liquid handling robotics and automated microscopy under its control, this learner accurately learned the effects of 48 chemical compounds on the subcellular localization of 48 proteins while performing only 29% of all possible experiments. The results represent the first practical demonstration of the utility of active learning-driven biological experimentation in which the set of possible phenotypes is unknown in advance.

Article and author information

Author details

  1. Armaghan W Naik

    Computational Biology Department, Center for Bioimage Informatics, Carnegie Mellon University, Pittsburgh, United States
    Competing interests
    The authors declare that no competing interests exist.
  2. Joshua D Kangas

    Computational Biology Department, Center for Bioimage Informatics, Carnegie Mellon University, Pittsburgh, United States
    Competing interests
    The authors declare that no competing interests exist.
  3. Devin P Sullivan

    Computational Biology Department, Center for Bioimage Informatics, Carnegie Mellon University, Pittsburgh, United States
    Competing interests
    The authors declare that no competing interests exist.
  4. Robert F Murphy

    Computational Biology Department, Center for Bioimage Informatics, Carnegie Mellon University, Pittsburgh, United States
    For correspondence
    murphy@cmu.edu
    Competing interests
    The authors declare that no competing interests exist.

Reviewing Editor

  1. Uwe Ohler, Duke, Germany

Version history

  1. Received: July 13, 2015
  2. Accepted: January 28, 2016
  3. Accepted Manuscript published: February 3, 2016 (version 1)
  4. Version of Record published: March 7, 2016 (version 2)

Copyright

© 2016, Naik et al.

This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 6,905
    views
  • 1,214
    downloads
  • 36
    citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Armaghan W Naik
  2. Joshua D Kangas
  3. Devin P Sullivan
  4. Robert F Murphy
(2016)
Active machine learning-driven experimentation to determine compound effects on protein patterns
eLife 5:e10047.
https://doi.org/10.7554/eLife.10047

Share this article

https://doi.org/10.7554/eLife.10047

Further reading

    1. Cell Biology
    2. Neuroscience
    Mariana I Tsap, Andriy S Yatsenko ... Halyna R Shcherbata
    Research Article

    Mutations in Drosophila Swiss Cheese (SWS) gene or its vertebrate orthologue Neuropathy Target Esterase (NTE) lead to progressive neuronal degeneration in flies and humans. Despite its enzymatic function as a phospholipase is well-established, the molecular mechanism responsible for maintaining nervous system integrity remains unclear. In this study, we found that NTE/SWS is present in surface glia that forms the blood-brain-barrier (BBB) and that NTE/SWS is important to maintain its structure and permeability. Importantly, BBB glia-specific expression of Drosophila NTE/SWS or human NTE in the sws mutant background fully rescues surface glial organization and partially restores BBB integrity, suggesting a conserved function of NTE/SWS. Interestingly, sws mutant glia showed abnormal organization of plasma membrane domains and tight junction rafts accompanied by the accumulation of lipid droplets, lysosomes, and multilamellar bodies. Since the observed cellular phenotypes closely resemble the characteristics described in a group of metabolic disorders known as lysosomal storage diseases (LSDs), our data established a novel connection between NTE/SWS and these conditions. We found that mutants with defective BBB exhibit elevated levels of fatty acids, which are precursors of eicosanoids and are involved in the inflammatory response. Also, as a consequence of a permeable BBB, several innate immunity factors are upregulated in an age-dependent manner, while BBB glia-specific expression of NTE/SWS normalizes inflammatory response. Treatment with anti-inflammatory agents prevents the abnormal architecture of the BBB, suggesting that inflammation contributes to the maintenance of a healthy brain barrier. Considering the link between a malfunctioning BBB and various neurodegenerative diseases, gaining a deeper understanding of the molecular mechanisms causing inflammation due to a defective BBB could help to promote the use of anti-inflammatory therapies for age-related neurodegeneration.

    1. Cell Biology
    2. Structural Biology and Molecular Biophysics
    Marcel Proske, Robert Janowski ... Dierk Niessing
    Research Article

    Mutations in the human PURA gene cause the neurodevelopmental PURA syndrome. In contrast to several other monogenetic disorders, almost all reported mutations in this nucleic acid-binding protein result in the full disease penetrance. In this study, we observed that patient mutations across PURA impair its previously reported co-localization with processing bodies. These mutations either destroyed the folding integrity, RNA binding, or dimerization of PURA. We also solved the crystal structures of the N- and C-terminal PUR domains of human PURA and combined them with molecular dynamics simulations and nuclear magnetic resonance measurements. The observed unusually high dynamics and structural promiscuity of PURA indicated that this protein is particularly susceptible to mutations impairing its structural integrity. It offers an explanation why even conservative mutations across PURA result in the full penetrance of symptoms in patients with PURA syndrome.