Improving prediction of phenotypic drug response on cancer cell lines using deep convolutional network
Related Experiment Video
Updated: Jan 21, 2026
Rapid in vivo Drug Response Prediction Using Leukemia Cell Grafts in Zebrafish Embryos
Published on: May 23, 2025
Pengfei Liu1, Hongjian Li2,3, Shuai Li4
1Department of Computer Science and Engineering, the Chinese University of Hong Kong, Sha Tin, N.T., Hong Kong, China. pfliu@cse.cuhk.edu.hk.
Background:
Understanding the phenotypic drug response on cancer cell lines plays a vital role in anti-cancer drug discovery and re-purposing. The Genomics of Drug Sensitivity in Cancer (GDSC) database provides open data for researchers in phenotypic screening to build and test their models. Previously, most research in these areas starts from the molecular fingerprints or physiochemical features of drugs, instead of their structures.
Results:
In this paper, a model called twin Convolutional Neural Network for drugs in SMILES format (tCNNS) is introduced for phenotypic screening. tCNNS uses a convolutional network to extract features for drugs from their simplified molecular input line entry specification (SMILES) format and uses another convolutional network to extract features for cancer cell lines from the genetic feature vectors respectively. After that, a fully connected network is used to predict the interaction between the drugs and the cancer cell lines. When the training set and the testing set are divided based on the interaction pairs between drugs and cell lines, tCNNS achieves 0.826, 0.831 for the mean and top quartile of the coefficient of determinant (R2) respectively and 0.909, 0.912 for the mean and top quartile of the Pearson correlation (Rp) respectively, which are significantly better than those of the previous works (Ammad-Ud-Din et al., J Chem Inf Model 54:2347-9, 2014), (Haider et al., PLoS ONE 10:0144490, 2015), (Menden et al., PLoS ONE 8:61318, 2013). However, when the training set and the testing set are divided exclusively based on drugs or cell lines, the performance of tCNNS decreases significantly and Rp and R2 drop to barely above 0.
Conclusions:
Our approach is able to predict the drug effects on cancer cell lines with high accuracy, and its performance remains stable with less but high-quality data, and with fewer features for the cancer cell lines. tCNNS can also solve the problem of outliers in other feature space. Besides achieving high scores in these statistical metrics, tCNNS also provides some insights into the phenotypic screening. However, the performance of tCNNS drops in the blind test.
Insights
A new twin Convolutional Neural Network (tCNN) model accurately predicts anti-cancer drug responses using drug structures and cell line genetic data. This approach advances drug discovery by analyzing phenotypic drug screening data more effectively.
Area of Science:
- Computational biology
- Bioinformatics
- Drug discovery
Background:
- Phenotypic drug response in cancer cell lines is crucial for anti-cancer drug discovery and repurposing.
- The Genomics of Drug Sensitivity in Cancer (GDSC) database offers valuable data for phenotypic screening models.
- Existing research often overlooks drug structures, focusing instead on molecular fingerprints or physiochemical features.
Purpose of the Study:
- To introduce a novel deep learning model, twin Convolutional Neural Network for drugs in SMILES format (tCNNS), for accurate phenotypic drug screening.
- To leverage drug structures (SMILES format) and cancer cell line genetic features for predicting drug-cancer interactions.
- To improve upon existing methods in anti-cancer drug discovery and repurposing through enhanced predictive accuracy.
Main Methods:
- The tCNNS model utilizes two parallel convolutional neural networks to extract features from drug SMILES strings and cancer cell line genetic data.
- A fully connected network integrates these features to predict drug-cancer cell line interactions.
- The model was evaluated on its ability to predict drug sensitivity based on interaction pairs, drugs, or cell lines.
Main Results:
- When data was split by interaction pairs, tCNNS achieved high performance with R² values of 0.826 (mean) and 0.831 (top quartile), and Pearson correlation (Rp) values of 0.909 (mean) and 0.912 (top quartile).
- These results significantly outperformed previous methods.
- Performance notably decreased when the training and testing sets were divided exclusively by drugs or cell lines, with R² and Rp dropping close to zero.
Conclusions:
- The tCNNS model demonstrates high accuracy in predicting drug effects on cancer cell lines, maintaining stability even with reduced, high-quality data and fewer features.
- The model can effectively handle outliers in feature spaces and offers insights into phenotypic screening.
- A limitation is the observed performance drop in blind tests, indicating areas for future improvement.
Related Concept Videos
02:26Protein Networks
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
01:17Convolution Properties II
The width property indicates that if the durations of input signals are T1 and T2, then the width of the output response equals the sum of both durations, irrespective of the shapes of the two functions. For instance, convolving two rectangular pulses with durations of 2 seconds and 1 second results in a function with a width of 3 seconds.
The area property asserts that the area under the...
01:16Cell Lines
01:20Convolution Properties I
The commutative property reveals that the input and the impulse response of an LTI (Linear Time-Invariant) system can be interchanged without affecting the output:
01:21Factors Affecting Drug Response: Overview
02:27Predicting Molecular Geometry