TRUST4 v1.1.4

• Fix several issues for Linux arm64 platform (#300 )
• Fix an issue of regarding genes like IGLL1 as constant genes for recent IMGT reference sequences.
• Add an option "--skipBarcodeErrorRead" option to fastq-extractor for customizable workflows (#301)

TRUST4 v1.1.3

• Add the option "--imgtAdditionalGap" to specify where IMGT introduces the additional gaps. This option is useful to handle the TRAV gene for mouse.
• Fix a serious bug in handling the additional gap that may cause the CDR3's 3' end shift when there are additional gaps. (#295 )
• Fix a serious bug that may crash the program when imputing TCR sequences (#294 )
• Fix a small bug of missing short V gene in a contig.

TRUST4 v1.1.2

• Detect special gaps in V genes from IMGT, e.g. mouse TRAV genes, and adjust the CDR3 coordinate accordingly.
• Always realign the reads to the contigs for abundance estimation. Orignally, this step is skipped in barcoded data. (#281 )
• Change the definition of the "complete_vdj" column as the sequence with 5' V to 3' J sequence. Originally, it requires the presence of C gene.
• Fix a bug in the airr_alignment file from annotator that may miss some column for contigs with CDR3 information
• Fix a bug of wrong V, J gene range after imputing TCRs

TRUST4 v1.1.1

• Improved assembly results for barcode-based data through barcode-wise k-mer count (#262)
• Output the imputed TCRs in the AIRR format for barcode data
• Fix a serious bug in smart-seq AIRR output. (#258)
• FIx a bug of getting wrong quality scores when using --readFormat

TRUST4 v1.1.0

• Change the license to MIT
• Significantly improved performance on data with barcode, e.g. 10x Genomics, TCR/BCR-seq with UMIs (--barcodeLevel molecule)
• Allow gzipped barcode whitelist
• Allow recursive folder creation by @TaverB (#252 )
• Add cell_id to smart-seq's AIRR-format output (#253 )
• Fix a serious bug that some reads may be used to calculate other barcode's chain's abundances (#254 )
• Fix a bug that the UMI information may be wrong when there are missing barcodes for 10x Genomics-like data (#254 )
• Fix a bug for some V and J gene alignment ranges if the anchor is too short

TRUST4 v1.0.14

• Improved sensitivity and precision by fixing a bug when merging two mate pairs.
• Add the option "--clean INT" to clean up intermediate files (#247)
• Do not suppress out-of-frame assemblies in "--barcodeLevel molecule" mode (#248)
• Allow TCR full-length assembly with --repseq option (#241 )
• Add c_cigar to the AIRR format output (#244)
• More robust implementation when there is no toassemble reads at all (#127)
• Suppress the CDR3s with length >= 7000, probably due to misannotation. (#242)
• Fix a bug if the read header is too long in the final.out file. (#237)
• Fix a bug where the "--barcode" option has to be specified before "--UMI". (#232)

TRUST4 v1.0.13

• "annotator" can take multiple-line fasta/fastq files now.
• The --read-format will not internally sort the segments, so users can fully control the read structure.
• Add "locus" column in the AIRR output.
• Fix a bug causing false positive TCRs during imputation. (An error was identified in the comment: https://academic.oup.com/bib/article/24/6/bbad354/7306821).
• Slightly improve the consensus choice when ties.

TRUST4 v1.0.12

• Fix a serious bug that may crash the program when generating AIRR alignment (#205 )
• Improve the robustness in testing example files (#206 )
• Improve the annotation accuracy when SHM is high
• Fix several false positive CDR3 issues
• Allow "/OR" genes in D gene annotation

TRUST4 v1.0.11

• "annotator" can directly output AIRR-format annotations, so it can be served as a stand-alone IgBLAST-like annotation method now. (https://github.com/liulab-dfci/TRUST4#annotation-only)
• Remove a redundant ';' in the secondary chain in the barcode_report file. (#197)

TRUST4 v1.0.10

• Add the option "--readFormat" that simplifies the read/barcode specification syntax.
• Add the option "--barcodeTranslate" to translate barcodes to another set of strings, and this option supports subset translation.
• Fix several issues in the perl script regarding creating IMGT sequence (Thanks to @Rudolph-afk )