LoRA-DR-suite is a family of models for the identification of disordered regions (DR) in proteins, built upon state-of-the-art Protein Language Models (PLMs) trained on protein sequences only. They leverage Low-Rank Adaptation (LoRA) fine-tuning for binary classification of intrinsic and soft disorder.
Intrinsically-disordered residues are experimentally detected through circular dichroism and X-ray cristallography, while soft disorder is characterized by high B-factor, or intermittently missing residues across different X-ray crystal structures of the same sequence.

Models for intrinsic disorder were evaluated on Critical Assessment of Intrinsic Disorder (CAID) data. In particular:

Models for soft disorder classification are trained instead on the SoftDis dataset, derived from an extensive analysis of clusters of alternative structures for the same protein sequence in the Protein Data Bank (PDB). For each position in the represantitive sequence of each cluster, the dataset provides the frequency of closely-related homologs for which the corresponding residue is higly flexible or missing. Any position with a frequency higher than 0 is labeled as soft disordered.

The data split for model training corresponds in particular to the id05 configuration, that further clusters representative sequences for each structure at 0.5 sequence identity. See the SoftDis Dataset card for more details.

Models trained with this dataset are denoted with the suffix “SD”.

We provide different model checkpoints, based on training data and pre-trained PLM.

* Union of DisProt7, CAID1 and CAID2 datasets

The models are intended to be used for classification of different disorder types.
Models for intrinsic disorder trained on DisProt 7.0 were evaluated on CAID1 and CAID2 challenge, but we suggest to use "ID" models for classification of new sequences, as they show better generalization, in particular esm2_650M-LoRA-ID.

In addition to its relation to flexibility and assembly pathways, soft disorder can be used to infer confidence score for structure prediciton tools, as we found high negative Spearman correlation between soft disorder probabilities and pLDDT from AlphaFold2 predicitons.

All models can be loaded as PyTorch Modules, together with their associated tokenizer, with the following code:

Once the model is loadded, disorder profile for all residues in a sequence can be obtained as follows:

@article{lombardi2025lora,
 title={LoRA-DR-suite: adapted embeddings predict intrinsic and soft disorder from protein sequences},
 author={Lombardi, Gianluca and Seoane, Beatriz and Carbone, Alessandra},
 journal={bioRxiv},
 year={2025},
}