Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.00143 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918312244412416 |
|---|---|
| author | Pan, Tom Dramko, Evan Miller, Mitchell D. Phillips Jr., George N. Kyrillidis, Anastasios |
| author_facet | Pan, Tom Dramko, Evan Miller, Mitchell D. Phillips Jr., George N. Kyrillidis, Anastasios |
| contents | Determining protein structures at an atomic level remains a significant challenge in structural biology. We introduce $\texttt{RecCrysFormer}$, a hybrid model that exploits the strengths of transformers with the aim of integrating experimental and ML approaches to protein structure determination from crystallographic data. $\texttt{RecCrysFormer}$ leverages Patterson maps and incorporates known standardized partial structures of amino acid residues to directly predict electron density maps, which are essential for constructing detailed atomic models through crystallographic refinement processes. $\texttt{RecCrysFormer}$ benefits from a ``recycling'' training regimen that iteratively incorporates results from crystallographic refinements and previous training runs as additional inputs in the form of template maps. Using a preliminary dataset of synthetic peptide fragments based on Protein Data Bank, $\texttt{RecCrysFormer}$ achieves good accuracy in structural predictions and shows robustness against variations in crystal parameters, such as unit cell dimensions and angles. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_00143 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RecCrysFormer: Refined Protein Structural Prediction from 3D Patterson Maps via Recycling Training Runs Pan, Tom Dramko, Evan Miller, Mitchell D. Phillips Jr., George N. Kyrillidis, Anastasios Quantitative Methods Machine Learning Optimization and Control I.2.1 Determining protein structures at an atomic level remains a significant challenge in structural biology. We introduce $\texttt{RecCrysFormer}$, a hybrid model that exploits the strengths of transformers with the aim of integrating experimental and ML approaches to protein structure determination from crystallographic data. $\texttt{RecCrysFormer}$ leverages Patterson maps and incorporates known standardized partial structures of amino acid residues to directly predict electron density maps, which are essential for constructing detailed atomic models through crystallographic refinement processes. $\texttt{RecCrysFormer}$ benefits from a ``recycling'' training regimen that iteratively incorporates results from crystallographic refinements and previous training runs as additional inputs in the form of template maps. Using a preliminary dataset of synthetic peptide fragments based on Protein Data Bank, $\texttt{RecCrysFormer}$ achieves good accuracy in structural predictions and shows robustness against variations in crystal parameters, such as unit cell dimensions and angles. |
| title | RecCrysFormer: Refined Protein Structural Prediction from 3D Patterson Maps via Recycling Training Runs |
| topic | Quantitative Methods Machine Learning Optimization and Control I.2.1 |
| url | https://arxiv.org/abs/2503.00143 |