Final Project: Mariam Ghochani – Statistical model in extracting protein associations

The anatomical distribution of multiple proteins within complex tissue is predicted to aid in the determination of musculodystrophy. The data from multiplexed proteomic imaging method of array tomography is a very rich data set that enables detection and characterization of each pixel with multiple numbers of markers. The tissue architecture of the samples and thus the inter relationship of multiple cells can be preserved and studied at resolutions that are higher than that of other conventional microscopy methods by a factor of 5-10. This essentially provides us with a rich high-dimensional dataset. Such dataset can be studied in order to characterize Changes in abundance, localization and distribution of single proteins. But also the pairwise relations from such dataset can be studied to characterize changes in the pairwise relationship of proteins in relation to the same complexes. The high dimensionality of this dataset will enable study of higher order and changes in the complex relationships of proteins as part of multi-partner ensembles as well as classification of compartments in terms of their associated heterogeneous protein markers. Having many markers per pixel will help to statistically increase the significance of classification of pairwise relationship, complex associations and classified compartments against sensitivity and specificity of recognition/detection. The underlying hypothesis for my model is that the false positive and false negative errors in the detection will not affect the pairwise and higher-order associations of proteins when analyzing/classifying with large number of markers and large distributions. For this the algorithm for modeling and simulation is as following:

1. Generate rich datasets.

2. Simulate different levels of noise in the dataset.

3. Confirm that with different levels of false positives and false negatives, the protein associations can still be extracted from dataset as long as the number of classifier and dataset is increased.

4. Prediction of the model: the sensitivity and robustness of the statistical model in extracting pairwise and higher order protein associations.

In this project, I attempted to develop a model for the proposed model-based statistical analysis that encompasses the alternate asymmetry of binding of DHPR to RyR in skeletal muscle.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s