By: Rohak Jain, Ma Lab (Seattle Children's Research Institute)
Project Aim: Drawing from the principle of synthetic lethality, our goal is to identify candidate host-directed intervention gene targets to tackle a broad range of viral infections. Through the computational analysis of existing gene expression data profiling Huh7 cells in baseline and dengue-infected conditions, we aim to (1) construct rigorous, condition-specific metabolic models for each treatment and (2) leverage these resources to predict synthetic lethal pairs with high confidence. Ultimately, we hope this investigation lays the groundwork for robust anti-viral therapies that target compromised host cells while leaving healthy bystanders unscathed.
![The general principle behind synthetic lethality (Image Credit: Xin & Zhang, 2023 [link])](https://images-provider.frontiersin.org/api/ipx/w=1200&f=png/https://www.frontiersin.org/files/Articles/1168143/fonc-13-1168143-HTML/image_m/fonc-13-1168143-g001.jpg)
The general principle behind synthetic lethality (Image Credit: Xin & Zhang, 2023; article).
To help you piece together the various elements of this research puzzle, the table below chronologically lists the intended purpose of each subsection, along with its main contents. This is by no means an exhaustive overview of every component, but it should function as a helpful guide for a first-time visitor!
| Section | Brief Description | Content Overview |
|---|---|---|
| (1) Input Data | Stores all files needed to execute workflows and scripts from start to finish. | Processed expression data (GSE110512), human GEM (sourced via SysBioCharmers), and relevant CRISPR screens. |
| (2) CBM Methods | Goes through all of the constraint-based metabolic modeling methods used to contextualize the dengue transcriptomic dataset. | End-to-end, annotated pipelines for running several CBM algorithms, including ones that are fully fleshed out (E-Flux, GIMME, RIPTiDe, tINIT) and others that are still in development (EXAMO, iMAT, mCADRE). |
| (3) Metabolomics Integration | Details the process of conducting 13C isotopically non-stationary MFA on the fluxomics measurements. | Data wrangling methods and core FreeFlux scripts. |
| (4) Downstream Analyses | Deals with the high-level secondary profiling of metabolic modeling results. | Statistically-based synthetic lethal prediction methods (Fast-SL, gMCS), visualizations in R, and much more. |
| (5) Project Documents | A place for presentations and additional materials (anything extraneous, really). | Project documents, primarily PowerPoints and literature reviews. |
Most of the development was conducted in Python, with versions ranging from 3.6 to 3.10. Because Jupyter notebooks and various tool-specific Conda virtual environments were used, please consult the setups.yml file to see a complete listing of package requirements for each environment. Once you've compiled a .yml suitable for your use-case, simply enter the following command into your terminal to ensure everything is arranged correctly:
conda env create --name envname --file environment.yml
Libraries:
- Python: cnapy, cobamp, cobra, copy, fastsl, freeflux, gmcspy, imatpy, mewpy, numpy, pandas, pymcadre, reframed, riptide, troppo.
- R: clusterProfiler, dplyr, ggforce, ggplot2, ggpubr, ggrepel, org.Hs.eg.db, patchwork, stringr.
Warning
This repo is a work in progress! All documents are subject to heavy revision as changes and improvements are made. Additionally, files prefixed by [INC] are incomplete / unfinished.
Please leave a comment in the 'Issues' tab if you have any questions or concerns. Thanks!