very low cross validated R2

[Yves33]

Hi (again),

After playing with my data (both glmnet and scikit-learn version), I find that my cross validated R2 are extremely low (<0.05) and even sometimes negative. How should I interpret this:
1 - there is definitively no relationship between my transcriptomes and my ephy properties
2 - the model fails to find this relationship
3 - there are some issues in my code / matrices

Thanks in advance
(see example more or less copied from demo.ipynb)