milaboratory/mixcr: MiXCR v4.0

Comprehensive support for Single-Cell and Molecular barcodes

• flexible and fast pattern matching engine to parse barcodes from the data; allows to fit the pipeline to any commercially available or in-house wet lab protocol with molecular or/and cell barcodes
• error correction in barcode sequences
• two cooperating UMI and/or Cell-barcode-based steps for clonal sequence reconstruction:
  • consensus assembly (i.e. for well-framed amplicon sequencing)
  • contig assembly (i.e. for 10x-like enzymatically fragmented data)
• tag information preserved on all analysis steps and extensive QC reports are generated throughout the pipeline, providing maximal visibility into analysis performance and giving a powerful tool for wet lab issues investigation

See the following usage examples:

• https://github.com/milaboratory/mixcr/wiki/Analysis-of-10X-Single-Cell-data
• https://github.com/milaboratory/mixcr/wiki/Analysis-of-full-length-BCR-data-with-unique-molecular-identifiers

Downstream analysis

Set of powerful downstream analysis features with the ability to export postanalysis results in tabular format and vector plots with various statistical comparisons.

• Ability to group samples by metadata values and compare repertoire features between groups
• Comprehensive repertoire normalization and filtering
• Statistical significance tests with proper p-value adjustment
• Repertoire overlap analysis
• Vector plots output (.svg / .pdf)
• Tabular outputs

See the following usage guide:

• https://github.com/milaboratory/mixcr/wiki/Downstream-analysis

Overlap browser

Added command exportClonesOverlap allowing to efficiently build and export overlap of the arbitrary number of clonesets.

Major rework of contig assembly algorithm

• significantly increased accuracy and stability
• works with or without molecular or cell barcodes
• can be applied to (sc)RNASeq data with reasonable IG/TCR coverage to reconstruct long sequence outside the CDR3

Export in AIRR format

• multiple options to export alignment or clonal data in AIRR format
• provides better compatibility with 3rd-party tools from AIRR community (see also RepSeq.IO feature for generation of fasta libraries with IMGT-like gaps from repseqio formatted references)

See here for usage example.

Other improvements and changes

• new built-in reference library with new species and newest genome based library for human (see changelog here)
• complete rewrite of IO for intermediate files (much faster IO with parallel serialization and deserialization, more compact files - each block is compressed with LZ4, versatile random access features provides additional speedup)
• faster hash-based external (file-based) sorting algorithm for alignment and other regrouping tasks in UMI/Single-cell related tasks and operations requiring alignment to clone mapping
• input sequence quality-score based trimming enabled by default
• support for human-readable alignments export from *.clna files by clone index
• all steps are cleaned-up to be completely pure, i.e. for the same input, output will always be byte-to-byte equal (no analysis date or other variable pieces of information leaks to the output files)
• more stable amino acid and combined amino acid plus nucleotide mutations export
• slight default analysis parameter optimization

Obtaining a license file

MiXCR requires a license file to run. Academic users with no commercial funding can quickly obtain a MiXCR license for free at https://licensing.milaboratories.com/. We are committed to support academic community and provide our software free of charge for scientists doing non-profit research. Commercial trial license can be requested at https://licensing.milaboratories.com or by email to licensing@milaboratories.com.

For details see: https://github.com/milaboratory/mixcr/wiki/Using-license