Very low and high values in metabolite to reaction matrix

[fratajcz]

Hi!

I just downloaded the Recon files from https://www.vmh.life/files/reconstructions/Recon/3D.01/Recon3D_301.zip and opened Recon3D_301.mat in python with scipy.io. However, I have some questions when looking at what I suppose is the metabolite to reaction matrix:

>>> recon[0]
<8399x13543 sparse matrix of type '<class 'numpy.float64'>'
        with 56791 stored elements in Compressed Sparse Column format>

>>> recon[0].A.sum()
82300.55691464

>>> recon[0].A
array([[-1.,  1.,  0., ...,  0.,  0.,  0.],
       [ 1.,  0.,  0., ...,  0.,  0.,  0.],
       [ 0., -1.,  0., ...,  0.,  0.,  0.],
       ...,
       [ 0.,  0.,  0., ...,  0.,  0.,  0.],
       [ 0.,  0.,  0., ...,  4.,  4.,  4.],
       [ 0.,  0.,  0., ...,  0.,  0.,  0.]])

>>> recon[0].A.min()
-18244.0

>>> recon[0].A.max()
77243.0

The questions are as follows:

• Why is the sum of the matrix not an integer or very close to an integer? Where does half a interaction come from?
• What do the values in there actually mean? I suppose negative values mean that this metabolite is an educt to the reaction, but is that so?
• How do you explain the extreme values of -18244 and 77243? That doesnt sound right to me.

If someone could point me towards the documentation of this file I would be very glad!
Thank you for your time and effort!