internship_coding

Repository for my M2 internship project:
Integrating long read data into Evolutionary Splicing Graphs (ESGs)
Supervisors: Elodie Laine (Sorbonne Université / Institut universitaire de France), Jean-Stéphane Varré (Université de Lille) and Arnaud Liehrmann (Sorbonne Université)

---

Overview

This project explores how long-read transcript predictions can be integrated into Evolutionary Splicing Graphs using the ThorAxe framework.

Key goals:

• Evaluate long-read transcript prediction tools (FLAIR, Bambu, IsoTools, RNA-Bloom)
• Develop a module to integrate custom transcript annotations into ThorAxe
• Assess tool complementarity and their impact on splicing event discovery

---

Main contributions

• add_transcripts.py: A Python module to add user-defined transcript structures to the ThorAxe input.
• Specification of a strict input CSV format derived from GTF and FASTA files.
• Comparative analysis of transcript-level and protein-level predictions from four tools.
• Integration of long-read–predicted isoforms into ESGs for human and mouse datasets.

---

Usage

Run the module from the command line:

python add_transcripts.py path/to/transcripts.csv path/to/gene/Ensembl