high unassigned alignments in clonotype assembly in a 10bp-UMI library

[ddzhang3]

Hello, I'm using mixcr v4.3.2 to analyze a customized TCR library, the command is like this:

mixcr analyze generic-tcr-amplicon-umi \
        --species mmu \
        --rna \
        --tag-pattern "^*(UMI:N{10})ATGGG(R1:*)\^(R2:*)" \
        --rigid-left-alignment-boundary \
        --floating-right-alignment-boundary C \
        --assemble-clonotypes-by VDJRegion \
        read1.fastq.gz read2.fastq.gz \
        result

However , the unassigned aligments is too high: 98%

Pre-clone assembler report:
Number of input groups: 4832
Number of input groups with no assembling feature: 41
Number of input alignments: 1718172
Number of alignments with assembling feature: 1467642 (85.42%)
Number of output pre-clones: 720
Number of pre-clonotypes per group:
0: + 4124 (86.08%) = 4124 (86.08%)
1: + 615 (12.84%) = 4739 (98.91%)
2: + 51 (1.06%) = 4790 (99.98%)
3: + 1 (0.02%) = 4791 (100%)
Number of assembling feature sequences in groups with zero pre-clonotypes: 1441481
Number of dropped pre-clones by tag suffix conflict: 0
Number of dropped alignments by tag suffix conflict: 0
Number of core alignments: 18994 (1.11%)
Discarded core alignments: 1448648 (7626.87%)
Empirically assigned alignments: 207 (0.01%)
Empirical assignment conflicts: 0 (0%)
Tag+VJ-gene empirically assigned alignments: 207 (0.01%)
VJ-gene empirically assigned alignments: 0 (0%)
Tag empirically assigned alignments: 0 (0%)
Number of ambiguous groups: 52
Number of ambiguous J-genes: 1
Number of ambiguous tag+V/J-gene combinations: 1
Ignored non-productive alignments: 0 (0%)
Unassigned alignments: 1697256 (98.78%)

my TCR library:

5'- AATGATACGGCGACCACCGAGATCTACAC-NNNNNNNN-ACACTCTTTCCCTACACGACGCTCTTCCGATCT-TSO-NNNNNNNNNN-ATGGG-5’UTR-signal peptide-V gene-（D gene）-J gene-part of C gene-AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC-NNNNNNNN-ATCTCGTATGCCGTCTTCTGCTTG -3'

the structure is:

5'-P5-index5-SP1-TSO-UMI-ATGGG-VDJ region-SP2-index7-P7-3'

the bold segment is the insert, and the sequencing mode is illumina PE300.

a snap of read1:

So, I was wondering if the --tag-pattern being setting wrong? Any advice?

[mizraelson]

Hi, the pattern appears to be correct. This issue may arise if you have many different clonotypes (distinct CDR3s) associated with the same UMI. Theoretically, only one or two molecules should correspond to a single UMI sequence.