NirsPredict 1.0


NirsPredict is an application allowing to make predictions about phenotypic traits of Arabidopsis thaliana by submiting related NIRS spectra using deep-learning. NirsPredict has three functionalities :

  • The submition of your data to our deep-learning algorithms to get traits predictions.
  • The consultation of the database used to construct our deep-learning models, it's possible to target the data you're interested in with some filters and download the corresponding dataset , morevover graphical representation of the dataset will be printed.
  • Contribute to the application by submiting your dataset so we can examine it and add it to our database/models

NirsPredict Application

Download application manual
Download metadata guide

Predict phenotypic traits

Submitted files must have headers

A dataset with at least 100 spectra is advised to get high-quality predictions

Download and consult the application manual above to get more details on input format and functionnalities


The prediction's robustness and the number of input to provide will rely on the selected mode.


Select some traits or/and metabolites to predict. Beware that the more you select the longer it will be.

You must provide a valid email adress before launch so you can receive your run's results
Prediction with our models can be imprecise due to specific conditions of your samples. In this case you can train against our deep learning scripts to build new model more suitable for your data to make predictions.

NirsPredict

NirsPredict is a tool allowing to make deep-learning predictions about phenotypic traits of A. thaliana by submiting related NIRS spectra

'A Perspective on Plant Phenomics: Coupling Deep Learning and Near-Infrared Spectroscopy' - Vasseur et al.

You can contact us on this email adress : NirsPredict@post.com

How to cite: Comming soon