Determine whether Scanpy's AnnData object has been logarithmized

1Just checking .rawwhether the attribute exists cannot accurately determine whether the AnnData object is logarithmic.

In this case, you can tell whether the AnnData object has been logarithmic by checking whether it has a non-zero negative value in its data matrix. If there are nonzero negative values ​​in the data matrix, you can infer that the AnnData object has been logarithmically transformed.

The following is a sample code that shows how to determine whether Scanpy's AnnData object has been logarithmized based on the data:

import numpy as np
 # 假设你的 AnnData 对象名为 adata 
if np.any(adata.X < 0): print("该 AnnData 对象已经被对数化") else: print("该 AnnData 对象未被对数化")

However, in rare cases in the log- log model, we often encounter negative values ​​after taking the logarithm of some variables , such as the ratio variable X. If 0<X<1, then after taking the logarithm , lnX will become negative value. Does this affect the regression results?