jbenbudd/ADPrLlama

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Mar 15, 2025Architecture:Transformer Cold

jbenbudd/ADPrLlama is a 7 billion parameter LoRA-fine-tuned model based on GreatCaptainNemo/ProLLaMA_Stage_1, designed for predicting ADP-ribosylation (ADPr) post-translational modification sites. It processes 21-residue peptide windows and outputs predicted sites in a structured format. This model specializes in biochemical sequence analysis, specifically for identifying ADPr sites.

Loading preview...

ADPrLlama: Predicting ADP-ribosylation Sites

ADPrLlama is a 7 billion parameter language model, specifically a LoRA-fine-tuned version of GreatCaptainNemo/ProLLaMA_Stage_1. Its primary function is to predict ADP-ribosylation (ADPr) post-translational modification (PTM) sites within protein sequences.

Key Capabilities

  • Specialized Prediction: Focuses exclusively on identifying ADPr PTM sites.
  • Input Format: Processes 21-residue peptide windows as input.
  • Output Format: Provides predictions in a structured format, e.g., Sites=<R5,D12,...>.
  • Biochemical Application: Tailored for specific tasks in bioinformatics and proteomics.

Good For

  • Biomedical Researchers: Ideal for scientists studying protein modifications and their biological implications.
  • Bioinformatics Pipelines: Can be integrated into workflows requiring automated ADPr site prediction.
  • Targeted Analysis: Useful for focused analysis of ADPr events in specific proteins or pathways.

This model is currently a training-only stub, with final performance metrics (ROC, accuracy, precision, recall, F1, confusion matrix) intended to be generated by a companion evaluation notebook on a held-out test set.