Fork me on GitHub


High-Throughput Analysis
of V(D)J Immune Repertoire


  • .fasta
  • .fastq
  • .gz
  • Patient database


  • Ultra-fast algorithm
  • 1 Gbp in < 5 minutes
    for one locus
  • On or self-hosted


  • Explore lymphocyte populations
  • Tag, annotate and merge clones
  • Send to IMGT/V-QUEST and to IgBlast
  • Generate a report
Vidjil is an open-source platform for the analysis of high-throughput sequencing data from lymphocytes, developed by the Bonsai bioinformatics team and its collaborators. V(D)J recombinations in lymphocytes are essential for immunological diversity. They are also useful markers of pathologies, and in leukemia, are used to quantify the minimal residual disease during patient follow-up. High-throughput sequencing (NGS/HTS) now enables the deep sequencing of a lymphoid population with dedicated Rep-Seq methods and software.

The Vidjil algorithm processes
     sequences recombined on the TRA/D, TRB, TRG, IGH, IGK and IGL locus,
     possibly with some incomplete/unusal recombinations such as Dh-Jh, Intron-KDE or Dd2-Dd3.

High-throughput algorithm

The Vidjil algorithm processes high-througput sequencing data to extract V(D)J junctions and gather them into clones. Vidjil starts from a set of reads and detects “windows” overlapping the actual CDR3. It detects gene rearrangements from both immunoglobulins and T-cell receptors, as well as some incomplete or uncommon rearrangements. The analysis is based on a reliable seed-based algorithm. It is extremely fast because, in the first phase, no alignment is performed with database germline sequences. The algorithm works on reads coming from either amplicon-based or capture-based deep sequencing strategy, as soon as they include CDR3 sequences. Both human and mouse immune systems can be analyzed.

Several views of the Vidjil web application displaying TRG clones of a patient with ALL: grid, time graph and clone sequences

Multi-sample web application

The Vidjil web application is made for the visualization, inspection and analysis of clones and their tracking along the time in a MRD setup or in a immunological study. The application can visualize data processed by the Vidjil algorithm or by other V(D)J analysis pipelines. It enables to explore further clusterings proposed by the software and/or done manually done by the user.

The patient view of the Vidjil web application. The database enables to store information on both the patient and individual samples.

Patient/experiment and sample database

The web application can be linked to a patient/experiment and sample database. After authentication, the clinicians or researchers upload .fasta/.fastq/.gz files, manage and process their runs directly from the web application. They can save their analysis and generate reports for the patient record. The server with the sample database can be installed in any hospital or computer center. A public web server is also available since October 2014. More than 20 labs in Europe and in the world regularly use Vidjil through the web application. In 18 months, the public server analyzed more than 5,000,000 sequences from 5,000 RepSeq samples.

Robust. Fast. Open-source.

Strong attention to users and stability. High-quality development process, with systemating testing and continous integration.

Extended collaborations

All the code of Vidjil is available under open-source licenses (GPLv3, as well as some other free licences for some third-parties librairies). We also offer extended support as well as custom development for various types of projects. Please contact us if you are interested.


Vidjil is developed by a passionate team with fast response, from the Bonsai bioinformatics team of the CRIStAL (CNRS, U. Lille) and Inria Lille research centers, in Lille, France. This work is in collaboration with the department of Hematology of CHRU Lille, the Functional and Structural Genomic Platform (U. Lille 2, IFR-114, IRCL), and the EuroClonality-NGS working group, and is supported by SIRIC ONCOLille (Grant INCa-DGOS-Inserm 6041), Région Nord-Pas-de-Calais (ABILES), Université Lille 1 (PPF Bioinformatique), and Inria Lille. The methods were presented at the JOBIM 2013 and ASH 2014 conferences, and are described in the following papers: