MeMDLM: De Novo Membrane Protein Design with Masked Discrete Diffusion Language Models

image/png

Masked Diffusion Language Models (MDLMs), introduced by Sahoo et al, provide strong generative capabilities to BERT-style models. In this work, we pre-train and fine-tune ESM-2-150M protein language model (pLM) on the MDLM objective to scaffold functional motifs and unconditionally generate realistic, high-quality membrane protein sequences.

Repository Authors

Shrey Goel, Undergraduate Student at Duke University
Vishrut Thoutam, Student at High Technology High School
Pranam Chatterjee, Assistant Professor at Duke University

Reach out to us with any questions!

Downloads last month
162
Safetensors
Model size
149M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for ChatterjeeLab/MeMDLM

Finetuned
(12)
this model