Negative results

[Li-Rui091]

Hi

I am using DWLS in my data, but I found the predicted proportions is negative, do you know why this happened?

[RLau0]

Hi Li-Rui091,

Did you find the solution to this? From the paper, it seems to suggest that it shouldn't be the case:
'given that cell-type proportions must be non-negative, the weighted least squares solution is constrained such that xj ≥ 0, for all j cell types.'