We are excited to announce that Cure Cancer With AI now offers access to the IBM MAMMAL (Molecular Aligned Multi-Modal Architecture and Language) model through our free public API. Developed by IBM Research, MAMMAL is a cutting-edge biomedical foundation model that plays a pivotal role in advancing cancer and drug-discovery research. This multi-modal model operates across various biological data types, including proteins, small molecules, and gene/omics data, making it a powerful tool for researchers and developers alike.

What is MAMMAL?

The term "Molecular Aligned Multi-Modal Architecture and Language" encapsulates the essence of this sophisticated model. MAMMAL integrates data from different biological modalities into a unified framework, representing proteins, small molecules, and transcriptomic data in a way that facilitates comprehensive analysis and predictions. This multi-task model has been trained on approximately 2 billion biological samples, allowing it to excel in various applications within the biomedical field.

The significance of a multi-modal biomedical foundation model like MAMMAL cannot be overstated. By incorporating diverse types of biological data, MAMMAL enables researchers to better understand complex interactions within biological systems, which is crucial for identifying potential drug candidates and advancing cancer treatments. In essence, it serves as a powerful AI-driven tool that enhances the drug discovery process, ultimately working towards better outcomes for patients.

The Research Behind MAMMAL

The capabilities of MAMMAL are backed by rigorous research, as documented in the paper “MAMMAL — Molecular Aligned Multi-Modal Architecture and Language for biomedical discovery,” published in npj Drug Discovery. You can read the full paper here. The research evaluated MAMMAL on 11 diverse drug-discovery tasks, achieving new state-of-the-art results on 9 of them. The performance of MAMMAL on the remaining two tasks was found to be comparable to existing state-of-the-art models, indicating its robustness and versatility.

With a training dataset of around 2 billion biological samples, MAMMAL's breadth of knowledge allows it to address a wide range of challenges in drug discovery, from predicting interactions between proteins to estimating the toxicity of drug candidates. This extensive training empowers researchers to leverage MAMMAL for various applications in their quest to develop effective cancer therapies.

MAMMAL in the Cure Cancer With AI API

We are thrilled to make MAMMAL's powerful predictions available for free through our public API. The API features three distinct endpoints that are designed to aid researchers in their drug discovery efforts:

• Protein–Protein Interaction Prediction (PPI)
  POST /api/v1/mammal/ppi
  This endpoint allows users to input two amino-acid sequences (protein_a and protein_b) and returns a binding-affinity class label. The output indicates whether the two proteins interact (1) or do not interact (0). This is essential for understanding biological pathways and mechanisms in cancer.
• Drug–Target Interaction Prediction (DTI)
  POST /api/v1/mammal/dti
  Users can provide a target protein's amino-acid sequence (target_seq) along with a drug in SMILES notation (drug_seq). The model predicts the pKd value (−log₁₀ Kd), with higher values indicating stronger predicted binding. This prediction is crucial for evaluating the potential efficacy of drug candidates against specific targets.
• ClinTox Clinical-Trial Toxicity Prediction
  POST /api/v1/mammal/clintox
  By entering a compound in SMILES notation (smiles), this endpoint provides a toxicity prediction, indicating whether the compound is toxic or likely to fail trials (pred 1 = toxic, 0 = not toxic) along with a raw score. This information can significantly influence drug development decisions.

Each of these endpoints is designed to provide valuable predictions that can guide research efforts in drug discovery. The predictions made by MAMMAL are for research purposes only and should not be considered medical or clinical advice.

Start Using It for Free

Accessing MAMMAL's capabilities is simple and free! Anyone can create an API key at /api-keys and start making requests. Our free tier allows up to 100 requests per hour per key, ensuring that researchers have the resources they need to explore this groundbreaking technology. To get started, you can also read the full documentation, including parameters and code samples, at /developers.

Conclusion

We are excited to offer the MAMMAL inference capabilities for free on curecancerwithai.com. This powerful biomedical foundation model can significantly enhance research in cancer and drug discovery, providing valuable insights into protein interactions, drug-target binding, and clinical toxicity. Remember, while MAMMAL provides critical predictions, these are research signals and not intended as medical advice. Begin your journey with MAMMAL today and unlock new possibilities in biomedical research!

To dive deeper into practical AI-for-cancer-research updates, explore our latest blog posts, learn more about our mission, and see how you can support ongoing work on our donations page.