Different alignment rates for MiXCR analyze and MiXCR align

[Januaryyiyue]

Hi,

I previously used MiXCR4.1.2 align, assemble, etc. separately and get an alignment log like this:

Total sequencing reads: 11689853
Successfully aligned reads: 10588551 (90.58%)
...

When I repeat the analysis with MiXCR analyze, my log file looks like this:

Total sequencing reads: 12150553
Successfully aligned reads: 202393 (1.67%)
...

The alignment rates are significantly different, but the resulting clones are almost the same.

Is there a reason why this is the case? If e.g., a previous alignment rate of 50% would be considered acceptable, what percentage should I expect from the new alignment rate?

Thank you!

[mizraelson]

Hi, can you share the exact commands?

[Januaryyiyue]

Sure.

This is the previous command:

module load mixcr/4.1.2
#SBATCH -J batch1_tcr_post_
java -jar /cluster/mixcr/4.1.2/mixcr.jar align -p rnaseq-cdr3 --verbose -s hsa -OallowPartialAlignments=true -OvParameters.geneFeatureToAlign=VGeneWithP -r /cluster/mixcr4/log/LOG_ALIGN_SAMPLE.txt /cluster/barcode_to_header/SAMPLE_barcode_R1.fastq /cluster/barcode_to_header/SAMPLE_barcode_R2.fastq /cluster/mixcr4/alignments/alignments_SAMPLE.vdjca
java -jar /cluster/mixcr/4.1.2/mixcr.jar assemblePartial /cluster/mixcr4/alignments/alignments_SAMPLE.vdjca /cluster/mixcr4/assemble/rescue1_SAMPLE.vdjca
java -jar /cluster/mixcr/4.1.2/mixcr.jar assemblePartial /cluster/mixcr4/assemble/rescue1_SAMPLE.vdjca /cluster/mixcr4/assemble/rescue2_SAMPLE.vdjca
java -jar /cluster/mixcr/4.1.2/mixcr.jar extend /cluster/mixcr4/assemble/rescue2_SAMPLE.vdjca /cluster/mixcr4/assemble/extended_SAMPLE.vdjca
java -jar /cluster/mixcr/4.1.2/mixcr.jar assemble -r /cluster/mixcr4/log/LOG_ASSEMBLE_SAMPLE.txt /cluster/mixcr4/assemble/extended_SAMPLE.vdjca /cluster/mixcr4/assemble/SAMPLE.clns
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones --chains TRA /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/clones/CLONES_TRA_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones --chains TRB /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/clones/CLONES_TRB_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones --chains TRD /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/clones/CLONES_TRD_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones --chains TRG /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/clones/CLONES_TRG_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones -o -t --chains TRA /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/productiveclones/CLONES_TRA_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones -o -t --chains TRB /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/productiveclones/CLONES_TRB_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones -o -t --chains TRD /cluster/mixcr4/assemble/SAMPLE.clns  /cluster/mixcr4/productiveclones/CLONES_TRD_SAMPLE.tsv
java -jar /cluster/mixcr/4.1.2/mixcr.jar exportClones -o -t --chains TRG /cluster/mixcr4/assemble/SAMPLE.clns /cluster/mixcr4/productiveclones/CLONES_TRG_SAMPLE.tsv

And this is the current command:

module load mixcr/4.6.0
mixcr -Xmx80g analyze local:cap-tcr /cluster/fastqlinks/SAMPLE_R1.fastq.gz /cluster/fastqlinks/SAMPLE_R2.fastq.gz /cluster/mixcr4.6.0_analyze/analyze_output/SAMPLE_mixcr_out

[mizraelson]

The difference is because you are using -OallowPartialAlignments=true, which preserves (marks as successfully aligned) reads that do not fully cover the CDR3 region. During the assembly step, such reads will not generate any clones due to the absence of the complete clonal sequence. This is particularly relevant for capTCR, where there are many reads resulting from the enrichment of the V and J genes that did not recombine. These reads will not contain the CDR3 region but will still be partially aligned. The analyze command filters out such reads on the align step.