Abstract
In many important contexts involving measurements of biological entities, there are distinct categories of information: some information is easy-to-obtain information (EI) and can be gathered on virtually every subject of interest, while other information is hard-to-obtain information (HI) and can only be gathered on some of the biological samples. For example, in the context of drug discovery, measurements like the chemical structure of a drug are EI, while measurements of the transcriptome of a cell population perturbed with the drug is HI. In the clinical context, basic health monitoring is EI because it is already being captured as part of other processes, while cellular measurements like flow cytometry or even ultimate patient outcome are HI. We propose building a model to make probabilistic predictions of HI from EI on the samples that have both kinds of measurements, which will allow us to generalize and predict the HI on a large set of samples from just the EI. To accomplish this, we present a conditional Generative Adversarial Network (cGAN) framework we call the Feature Mapping GAN (FMGAN). By using the EI as conditions to map to the HI, we demonstrate that FMGAN can accurately predict the HI, with heterogeneity in cases of distributions of HI from EI. We show that FMGAN is flexible in that it can learn rich and complex mappings from EI to HI, and can take into account manifold structure in the EI space where available. We demonstrate this in a variety of contexts including generating RNA sequencing results on cell lines subjected to drug perturbations using drug chemical structure, and generating clinical outcomes from patient lab measurements. Most notably, we are able to generate synthetic flow cytometry data from clinical variables on a cohort of COVID-19 patients—effectively describing their immune response in great detail, and showcasing the power of generating expensive FACS data from ubiquitously available patient monitoring data.
The Data Bottleneck in Biomedicine
In biomedical research, a persistent "data bottleneck" exists between the volume of data we can collect and the depth of insight that data provides. We can easily collect "Easy-to-obtain Information" (EI)—such as the chemical formula of a drug or a patient’s heart rate—on millions of subjects. However, "Hard-to-obtain Information" (HI)—like how that drug changes the expression of 20,000 genes or the specific state of a patient’s immune cells—is expensive, time-consuming, and often invasive.
The challenge is that the relationship between EI and HI is rarely linear or deterministic. A tiny change in a drug's chemical structure can lead to a massive change in cellular response. Furthermore, biological systems are inherently noisy; repeating the same experiment can yield slightly different results. Traditional regression models struggle with these "one-to-many" mappings and often fail to capture the complex, manifold structure of biological data.
The Feature Mapping GAN (FMGAN) Framework
To address this, we introduce the Feature Mapping GAN (FMGAN), based on a Conditional Generative Adversarial Network (cGAN) architecture. Unlike a standard GAN that generates images from random noise, a cGAN generates data based on a "condition." In FMGAN, the EI serves as the condition, and the HI is the target data distribution.
Handling Complexity and Manifold Learning
To make diverse EI sources usable, FMGAN employs specialized encoders. For drug discovery, we use Convolutional Neural Networks (CNNs) to process either SMILES strings (textual representations of chemicals) or 2D structure diagrams. This creates a regularized latent space where biological distances are preserved.
In cases where the EI is already numerical but high-dimensional, we utilize PHATE (a dimensionality reduction tool) to find the "intrinsic coordinates" of the data. This helps the GAN understand the underlying biological "shape" of the input space, making the mapping to HI much more accurate.
Experimental Validation and Results
We validated FMGAN across three distinct domains using Maximum Mean Discrepancy (MMD) to evaluate how closely the generated distributions matched real experimental data (lower is better):
- Drug Perturbations (L1000 Connectivity Map): FMGAN significantly outperformed linear baselines. When predicting gene expression from chemical images, FMGAN achieved an MMD of 1.177 compared to the baseline's 1.798.
- Clinical Mortality (eICU Database): In predicting patient mortality across the patient manifold, FMGAN achieved an error rate below 0.01, while linear baseline errors were 2x to 10x higher.
- COVID-19 FACS Generation: This was the most striking result. FMGAN generated synthetic flow cytometry (FACS) data from simple blood work that closely matched real patient distributions (MMD 0.022 vs. baseline 0.898). The model successfully identified specific T-cell populations (CD3+, CD4+, CD8+) using only routine clinical monitoring data.
Discussion and Future Directions
FMGAN’s ability to generate a range of outcomes makes it more biologically "honest" than deterministic models, treating stochasticity as a feature rather than a bug. By simulating experimental replicates, it provides a realistic distribution of potential biological responses.
However, the model’s accuracy depends on the representativeness of the training set. If the training data doesn't capture a specific biological mechanism, the model cannot accurately predict it for new samples. Future work will focus on integrating more diverse EI sources, such as unstructured doctor’s notes using Natural Language Processing (NLP), and applying the framework to other expensive modalities like high-resolution medical imaging. This approach has the potential to re-design pharmaceutical pipelines and provide "virtual diagnostics" in clinical settings.
Summary updated