Convolutional neural network-based regression analysis to predict subnuclear chromatin organization from...

Azimuth-resolved optical scattering signals obtained from cell nuclei are sensitive to changes in their internal refractive index profile. These two-dimensional signals can therefore offer significant insights into chromatin organization.

We aim to determine whether two-dimensional scattering signals can be used in an inverse scheme to extract the spatial correlation length ${\ell }_c$ and extent $\delta n$ of subnuclear refractive index fluctuations to provide quantitative information on chromatin distribution.

Since an analytical formulation that links azimuth-resolved signals to ${\ell }_c$ and $\delta n$ is not feasible, we set out to assess the potential of machine learning to predict these parameters via a data-driven approach. We carry out a convolutional neural network (CNN)-based regression analysis on 198 numerically computed signals for nuclear models constructed with ${\ell }_c$ varying in steps of $0.1\mu m$ between 0.4 and $1.0\mu m$ , and $\delta n$ varying in steps of 0.005 between 0.005 and 0.035. We quantify the performance of our analysis using a five-fold cross-validation technique.

The results show agreement between the true and predicted values for both ${\ell }_c$ and $\delta n$ , with mean absolute percent errors of 8.5% and 13.5%, respectively. These errors are smaller than the minimum percent increment between successive values for respective parameters characterizing the constructed models and thus signify an extremely good prediction performance over the range of interest.

Our results reveal that CNN-based regression can be a powerful approach for exploiting the information content of two-dimensional optical scattering signals and hence monitoring chromatin organization in a quantitative manner.