# Knowledge Discovery in Databases of Proteomics by Systems Modeling in Translational Research on Pancreatic Cancer

## TL;DR

This paper introduces a systems modeling workflow to discover translational targets in pancreatic cancer by integrating proteomics data and bioinformatics.

## Contribution

The novel contribution is a systems modeling framework for KDD that identifies translational targets in pancreatic cancer using proteomics and AI.

## Key findings

- Common proteins were identified between human PDAC and in vitro/in vivo models.
- Potential translational targets were hypothesized based on hub proteins and signaling pathways.
- PDAC-specific proteins and high topological proteins were highlighted as key findings.

## Abstract

Background: Knowledge discovery in databases (KDD) can contribute to translational research, also known as translational medicine, by bridging the gap between in vitro and in vivo studies, and clinical applications. Here, we propose a ‘systems modeling’ workflow for KDD. Methods: This framework includes the data collection of a composition model (various research models), processing model (proteomics) and analytical model (bioinformatics, artificial intelligence/machine leaning and pattern evaluation), knowledge presentation, and feedback loops for hypothesis generation and validation. We applied this workflow to study pancreatic ductal adenocarcinoma (PDAC). Results: We identified the common proteins between human PDAC and various research models in vitro (cells, spheroids and organoids) and in vivo (mouse mice). Accordingly, we hypothesized potential translational targets on hub proteins and the related signaling pathways, PDAC-specific proteins and signature pathways, and high topological proteins. Conclusions: This systems modeling workflow can be a valuable method for KDD, facilitating knowledge discovery in translational targets in general, and in particular to PADA in this case.