| Accession |
PRJCA028510 |
| Title |
PEAKFORMER: Transformer-based Precise Peak Identification and Quantification Tool |
| Relevance |
Medical |
| Data types |
metabolomics
|
| Organisms |
Homo sapiens
|
| Description |
We developed PeakFormer, a deep learning method based on object detection designed to detect complete peak signals. Our algorithm harnesses the capabilities of transformers, training on over 20,000 annotated EIC and ensuring unique predictions through bipartite matching. Without retraining, PeakFormer achieves over 90% accuracy in distinguishing true and false peaks on the EVA dataset. We performed interpretability analysis of the encoder and decoder using visualization techniques. |
| Sample scope |
clinical cohorts |
| Release date |
2024-07-25 |
| Grants |
| Agency |
program |
Grant ID |
Grant title |
| National Natural Science Foundation of China (NSFC)
|
Major Research Plan
|
91957120
|
Based on spalial melabolomics, the metabolic mode of aldose reductase inpromoting the occurence and development of liver cancer was investigated
|
| National Natural Science Foundation of China (NSFC)
|
General Program
|
21974114
|
Mass specrometry-based immunometabolic landscape analysis of lung metastases in breast cancer
|
|
|
| Submitter |
Shuhai
Lin (shuhai@xmu.edu.cn)
|
| Organization |
Xiamen University |
| Submission date |
2024-07-25 |
