OLINK Proteomic Analysis to Identify Potential Extracellular Vesicle-Associated Proteins

Analysis of OLINK proteomic data to identify proteins that may be associated with brain-derived extracellular vesicles.

Key Features

• A dataset containing information concenrning 5416 unique proteins, collected via the OLINK HT panel using frationated human cerebrospinal fluid.
• Read OLINK parquet files and identify proteins that may be associated with extracellular vesicles using relative protein abundances in fractionated human cerebrospinal fluid.
• Overlay lists of proteins that may be associated with extracellular vesicles with single-cell RNA sequencing data and subcellular localization analysis to determine if a particular protein could be a potential cell-type specific immunocapture or validation target.

Modules

config.py

• Contains several global variables.

raw_data_preprocessing.py

• Converts the raw parquet file produced by OLINK into a tidy dataframe.
• Generates graphs to display the median fractionation pattern of a protein of interest.
• Calculates the EV Association Score of a protein of interest.

Required Packages

• matplotlib.axes
• matplotlib.pyplot
• pandas

Required Documentation

• config.py

olink_fractionation.py

• Uses fractionation patterns reported by Olink to identify proteins that may be associated with extracellular vesicles.

Required Packages

• pandas

specificity_functions.py

• Calculates various statistical measures of specificity, including tau score, tissue specificity index, gini coefficient, Shannon entropy, specificity measure, and zscore.

Required Packages

• numpy
• pandas
• scipy

brainrnaseq_specificity.py

• Uses data collected and made available by BrainRNA-Seq to determine proteins that are specific to a cell type of interest.

Required Packages

• requests
• numpy
• pandas
• io
• pathlib

Required Documentation

• config
• specificity_functions

gtex_specificity.py

• Uses data made available by GTEx to determine proteins that are specific to the brain.

Required Packages

• pandas

Required Documentation

• specificity_functions
• config

deeptmhmm_localization.py

• Uses the DeepTMHMM deep learning model to identify the most likely subcellular localization of proteins of interest.

Required Packages

• biolib
• gzip
• os
• pathlib
• pandas

identify_targets.py

• Uses specified fractionation, localization, and cell-type specificity criteria to identify protein targets.

Required Packages

• pandas
• typing
• pathlib

Required Documentation

• raw_data_preprocessing.py
• olink_fractionation.py
• brainrnaseq_specificity.py
• deeptmhmm_localization.py
• config.py