Motivation

DeepPROTECTNeo is a comprehensive web server designed to streamline neoantigen discovery for personalized cancer vaccines. By processing Whole Exome Sequencing (WES) or Whole Genome Sequencing (WGS) data, it integrates HLA typing, TCR profiling, peptide-MHC binding prediction, and TCR-epitope interaction assessment into a single workflow. Powered by advanced deep learning, including context-aware transformers, it achieves unparalleled precision in neoepitope discovery and supports multi-epitope vaccine design. With extensive validation on clinical datasets like TESLA, DeepPROTECTNeo offers researchers and clinicians an innovative tool to advance cancer immunotherapy.

Process Flow Steps

  1. We perform quality control and trimming on the raw FASTQ file provided by the users
  2. Alignment and Post-processing is done on the processed read files
  3. Variant calling and annotation is performed to detect somatic mutations
  4. We identify HLA types from the processed BAMs
  5. We perform peptide-MHC binding affinity prediction
  6. We also identify TCR (CDR3-beta) regions from the given input data
  7. Perform TCR-peptide binding
  8. Peptides having high TCR affinity are selected for BCR affinity Prediction to produce the final result

User Guidelines

Strictly follow the file formats and naming conventions provided here to get your result :

  • Input paired-end FASTQ files should be named as: example_1.fastq.gz, example_2.fastq.gz
  • Input BAM file should be named as: example.bam
  • For TCR-peptide bindings, input should be a txt file with 2 columns, peptide followed by TCR regions separated by a tab. No headers required.
  • Find example files to know about data format and naming convention in our Help page.
Process Flow Diagram

DeepPROTECTNeo accepts files of size <= 3GB. For larger files, please email (debrajccds@kgpian.iitkgp.ac.in) your data with the subject "DeepPROTECTNeo Job Request". We will respond as soon as the job completes.