We validated the presented model of WNT/fi-catenin signaling against independent in-silico and in-Vitro data [27, 49].

Here we apply a combined in-vitro and in-silico approach to investigate the spatio-temporal regulation of WNT/[3-catenin signaling during the early neural differentiation process of human neural progenitors cells (hNPCs), which form a new prospect for replacement therapies in the context of neuro-degenerative diseases.

The model’s predictive ability is demonstrated under a wide range of varying conditions for in-vitro and in-silico reference data sets.

Human neural progenitor cells offer the promising perspective of using in-vitro grown neural cell populations for replacement therapies in the context of neurodegenerative diseases, such as Parkinson’s or Huntington’s disease.

Based on a combined in-vitro and in-silico approach we demonstrate an elaborate interplay between endogenous ROS and lipid raft dependent WNT/beta-catenin signaling controlling the nuclear beta-catenin levels throughout the initial phase of neural differentiation.

In a combined in-vitro and in-silico approach we find strong evidence, that cell fate commitment in human neural progenitor cells is driven by two distinct fi-catenin signaling mechanisms.

NPCs provide a new, promising basis for the in-vitro growth of neuron populations that can be used in replacement therapies for neurodegenerative diseases, such as Parkinson’s or Huntington’s diseases [6, 7].

We use literature values as often as possible and fit the remaining parameters to experimental measurements of nuclear fi-catenin dynamics during in-vitro differentiation of ReNcell VM 197 cells.

To further test the calibrated/fitted model we apply cross-validation by reproducing existing in-silico and in-vitro data (measurements of fi-catenin concentration under different WNT stimuli).

Therefore we investigate the mutual influence of lipid rafts on WNT-signaling during the in-vitro differentiation of immortalized human neural progenitor cells (ReNcell VM197).

The model is based on experimental data as well as literature values and has been extensively validated against in-vitro and in-silico data under a Wide range of varying conditions.

The mean and standard deviations of DMSO and Taxol controls for the in-vitro tubulin polymerization assays were calculated and used to scale the compound OD readout between different runs to normalize the heterogeneity of the reaction.

All the statistical analysis for in-vitro tubulin polymerization assays was performed using Microsoft Excel.

(A) 212 antimitotic compounds clustered into 85 distinct chemical similarity sub-networks of which 23 clusters contained annotated anti-tubulin agents (green); additionally 54 novel tubulin-targeting chemotypes (yellow) were identified from in-vitro tubulin polymerization assays.

Of the 51 compounds predicted to be targeting microtubules, 36 compounds (71%) had more than 20% fold change in in-vitro tubulin polymerization assay and 14 had no measurable effect.

(B) Tubulin polymerization kinetics for 7 novel tubulin destabilizers (6—12), based on a phenyl-sulfanyl-thiazol-acetamide scaffold, using an in-vitro tubulin polymerization assay.

Based on target prediction, we selected microtubules (0c and B-tubulin) as our target for in-vitro validation.

To test CSNAP’s prediction that 51 of the 212 mitotic compounds were targeting microtubules, we reacquired all 212 compounds and tested their ability to perturb micro-tubule polymerization (stabilize or destabilize microtubules) in an in-vitro microtubule polymerization assay at 50uM concentration (Fig.

In addition, in-vitro testing led to the discovery of 96 additional compounds for a total of 132 anti-tubulin agents, including structurally diverse compounds covering ~54 novel chemotypes not discovered in previous chemical screens (S3 Table).

Machine learning techniques may replace expensive in-vitro laboratory experiments by learning an accurate model of it.

Finally, in-vitro and in-silico results are provided to support and validate this theoretical discovery.

I n-silico predictions are faster and cheaper than in-vitro assays, however, predicting the bioactivity of all possible peptide to select the most bioactive ones would require a prohibitive amount of computational time.

Comparing hmndom and the oracle accuracies on the CAMPs and BPPs databases To provide additional support for its accuracy, predictor hmndom was used to predict the bioactivity values of unseen but in-vitro validated peptides of the CAMPs and BPPs databases.