Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Is Multitask Deep Learning Practical for Pharma?

Bharath Ramsundar1, Bowen Liu2, Zhenqin Wu2

  • 1Department of Computer Science, Stanford University , Stanford, California 94305, United States.

Journal of Chemical Information and Modeling
|July 11, 2017
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Drug Discovery: Overview01:26

Drug Discovery: Overview

12.3K
Drug discovery is a multifaceted process involving extensive screening, testing, and optimization of lead compounds to identify potential new drugs for therapeutic use. It combines several approaches, including screening large numbers of natural products, chemical modification of known active molecules, identification of new drug targets, and rational design based on biological mechanisms and drug-receptor structure. These approaches are carried out in both academic research laboratories and...
12.3K
Preclinical Development: Overview01:28

Preclinical Development: Overview

6.1K
Preclinical development consists of a series of tests that ensure the safety and efficacy of a new therapeutic compound before it is tested in humans. There are four main phases to this process. First, safety pharmacology tests are conducted to ensure the drug does not produce any acutely harmful effects. These tests examine parameters such as bronchoconstriction, cardiac dysrhythmias, blood pressure changes, and ataxia. Next, preliminary toxicological testing is performed to determine the...
6.1K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Structure-Based Discovery of Imidazo[4,5-<i>c</i>]pyridine SARM1 Modulators Showing Paradoxical Activation.

Journal of medicinal chemistry·2026
Same author

Gungnir codec enabling high error-tolerance and low-redundancy DNA storage through substantial computing power.

Nature communications·2026
Same author

NS5-targeting nucleoside analogs inhibit dengue virus and other flaviviruses.

PLoS pathogens·2026
Same author

Enhancing Permeability Prediction of Heterobifunctional Degraders Using Machine Learning and Metadynamics-Informed 3D Molecular Descriptors.

Journal of chemical information and modeling·2025
Same author

Discovery of a Novel Series of <i>iso</i>-Indolinone-Based Glutarimides as Highly Efficacious and Selective IKZF2 Molecular Glue Degraders.

Journal of medicinal chemistry·2025
Same author

ROSIE: AI generation of multiplex immunofluorescence staining from histopathology images.

Nature communications·2025
Same journal

Correction to "AstraMEV (AI-Guided Structural Assembly of Multi-Epitope Vaccines) Against Infectious Bronchitis Virus".

Journal of chemical information and modeling·2026
Same journal

MolPy: A Large Language Model-Friendly Toolkit for Reactive Topology Editing in Polymer Simulations.

Journal of chemical information and modeling·2026
Same journal

Molecular Mechanisms of KIT Receptor Dimerization and Oncogenic Activation Revealed by Multiscale Simulations.

Journal of chemical information and modeling·2026
Same journal

Structural and Thermodynamic Discrimination between Agonists and Antagonists of Retinoic Acid Receptor γ and the Vitamin D Receptor.

Journal of chemical information and modeling·2026
Same journal

PACEff Builder: An Efficient Platform for Constructing PACE Hybrid-Resolution Models for Molecular Dynamics Simulations of Aqueous Protein, Peptide Assembly, and Membrane Protein Systems.

Journal of chemical information and modeling·2026
Same journal

TransKla: A Local-Global Cross-Attention Based Transformer Approach for Prediction of Lysine Lactylation Sites.

Journal of chemical information and modeling·2026
See all related articles

Multitask deep learning models are now ready for drug discovery. An open-source tool in DeepChem demonstrates their robustness and improved performance over traditional methods.

Area of Science:

  • Computational chemistry
  • Machine learning in drug discovery

Background:

  • Multitask deep learning (MTDL) shows promise for drug discovery but faces adoption barriers.
  • Lack of robust software implementations and understanding of MTDL performance hinders industry use.

Purpose of the Study:

  • To address software and understanding barriers for MTDL adoption in drug discovery.
  • To provide an open-source implementation of MTDL within the DeepChem platform.
  • To analyze MTDL performance on pharmaceutical datasets under challenging conditions.

Main Methods:

  • Developed an open-source MTDL implementation in DeepChem for model construction, fitting, and evaluation.
  • Analyzed MTDL performance on four pharmaceutical datasets using time and neighbor splits for rigorous testing.

Related Experiment Videos

  • Compared MTDL performance against random forests and other deep learning models.

    • Main Results:

    • MTDL models demonstrated surprising robustness across pharmaceutical datasets.
    • MTDL significantly outperformed random forests in computational drug discovery tasks.
    • The open-source implementation facilitated straightforward analysis and evaluation of MTDL models.

    Conclusions:

    • MTDL models are robust and offer significant advantages over traditional methods for drug discovery.
    • The DeepChem open-source implementation lowers barriers to MTDL adoption in the pharmaceutical industry.
    • MTDL is ready for widespread application in commercial drug discovery pipelines.