How to choose the best preset for certain data?

[linqy-immune]

Hi, @mizraelson
I'm here again. Please receive my sincere thanks for your kind help and the maintenance for the wonderful and great software.
I'm very sorry that although I have used MiXCR for almost 1 year, I haven't had mastered the principles and usages well. Recently it hits on me that wheteher it is right that I chose the analysis preset by the data sequenced platform and sometimes the kit before. But it seems that no all the analysis outcomes are good and I think maybe it is resulted from the wrong preset. Could you please tell me how to choose the preset reasonably?
There is an example, for the fastq data in PRJNA1147802, the Instrument was Illumina MiSeq, Strategy was AMPLICON, Source was TRANSCRIPTOMIC, Selection was PCR, Layout was SINGLE but the data title shows AIRR-seq, whether it is proper for me to choose "illumina-human-rna-trb-ampliseq-plus" preset? If not, which preset do you recommend?
I am looking forward to your reply.
Thanks and best regards,
Linqy

[linqy-immune]

Hi, @mizraelson
What preset should I set for the data from E-MTAB-11498? It is shown that Total RNA was used for cDNA synthesis using 5’RACE template switch technology to introduce universal primer binding site, unique molecular identifiers (UMI) and first sample barcode at the 5’ end of RNA molecules. Primers complementary to TCRalpha and TCRbeta constant segments were used for initiation of cDNA synthesis. cDNA was amplified in two subsequent PCR steps. During the second PCR step, second sample barcode and adapters were introduced to the libraries. The libraries were sequenced on NextSeq, HiSeq2000/2500 or MiSeq (Illumina) with 2x100bp or 2x150bp sequencing length. in the article TCR repertoire profiling revealed antigen-driven CD8+ T cell clonal groups shared in synovial fluid of patients with spondyloarthritis. I'm not sure whether the preset mixcr analyze -s hs milab-human-rna-tcr-umi-race I use is right cause I find that the export outcomes only with 11 kB.
Could you please give me some tips? I really need your help!!!

[mizraelson]

Hi, the choice of preset depends on the kit used to generate the cDNA library. For example, milab-human-rna-tcr-umi-race should only be used for data generated with the MiLaboratories Human TCR 5’ RACE kit. The preset illumina-human-rna-trb-ampliseq-plus should only be used for cDNA libraries generated with the Illumina Ampliseq Plus Human TCR beta kit.

For more generic or in-house protocols, you can use the generic-amplicon or generic-amplicon-with-umi presets, depending on whether the cDNA library includes molecular barcodes or not. The most generic command for non-UMI data would be:

mixcr analyze generic-amplicon \
    --species mmu \
    --rna \
    --floating-left-alignment-boundary \
    --floating-right-alignment-boundary C \
    --assemble-clonotypes-by CDR3 \
      input_R1.fastq.gz \
      input_R2.fastq.gz \
      result

Make sure to select the correct species.

[linqy-immune]

Got it! Thank you for your explanation! It is very kind of you!
But I am still a little confused about which preset to use for the data Libraries of TCRalpha and TCRbeta chains were prepared with the protocol that implements double barcoding of samples and labeling of cDNA with unique molecular identifiers, allowing for potential cross-sample contamination removal, error-correction, and data normalization [described in (22)]. In brief, total RNA from MNCs was isolated using TRIzol reagent (Thermo Fisher Scientific) or RNeasy Mini kit (Qiagen). Total RNA was used for cDNA synthesis using 5’RACE template switch technology to introduce universal primer binding site, unique molecular identifiers (UMI) and first sample barcode at the 5’ end of RNA molecules. Primers complementary to TCRalpha and TCRbeta constant segments were used for initiation of cDNA synthesis. cDNA was amplified in two subsequent PCR steps. During the second PCR step, second sample barcode and adapters were introduced to the libraries. Libraries of TCRalpha and TCRbeta chains were prepared with the protocol that implements double barcoding of samples and labeling of cDNA with unique molecular identifiers, allowing for potential cross-sample contamination removal, error-correction, and data normalization [described in (22)]. In brief, total RNA from MNCs was isolated using TRIzol reagent (Thermo Fisher Scientific) or RNeasy Mini kit (Qiagen). Total RNA was used for cDNA synthesis using 5’RACE template switch technology to introduce universal primer binding site, unique molecular identifiers (UMI) and first sample barcode at the 5’ end of RNA molecules. Primers complementary to TCRalpha and TCRbeta constant segments were used for initiation of cDNA synthesis. cDNA was amplified in two subsequent PCR steps. During the second PCR step, second sample barcode and adapters were introduced to the libraries.? Do you mean that it is better to use generic-amplicon-with-umi preset without considering the sequence platform?
Is it right that almost all single-cell TCR sequence data should be analyzed by 10x-sc-xcr-vdj preset?
On the other hand, I am not very sure what floating-left-alignment-boundary means? When and where should I use this parameter?
Looking forward to your reply and thank you again for your selfless help!
Linqy

[mizraelson]

The sequencing platform generally doesn’t matter unless it’s ONT or PacBio. For the protocol you described, you should use generic-amplicon-with-umi and provide the UMI structure using the --tag-pattern parameter.

Only 10x 5’ VDJ data should be analyzed with the 10x-sc-xcr-vdj preset. The choice of preset depends entirely on the library structure, not the sequencing platform.

floating-left-alignment-boundary :

• Configures aligners to use semi-local alignment at reads 5'-end. Typically used with V gene single primer / multiplex protocols, or if there are non-trimmed adapter sequences at 5'-end. Optional anchor point may be specified to instruct MiXCR where the primer is located and strip V feature to align accordingly, resulting in a more precise alignments.

[linqy-immune]

Hello! @mizraelson
I'm here again asking for help. Recently, I try to analyze single-cell data from the downloaded sra files. But something weird happens when I transfer the sra files into fastq files: there are 4 fastq files for most of the same batch sra files, but for another several sra files, only 3 fastq files are got. The extraction protocol is scV(D)J-seq: TCR and BCR sequencing was performed using cDNA synthesized by the nested-PCR enrichment method. TCR and BCR libraries were generated using the Chromium Single Cell Human TCR and BCR Amplification Kit (10X Genomics), and each library was checked for quality control (Agilent Bioanalyzer) before sequencing.

Can I analyze this batch of files by MiXCR? Which files should I choose? Could you please give me some instructions? If convenient for you, could you recommend the code for mixcr analysis?
Thanks again for your kind help and maintenance.

[mizraelson]

Judging by the length, I think the reads that are 10 nt long are the index reads (I1 and I2). The one that is 150 bp is probably R2. However, I don’t see R1 - it should be at least 26 bp long for 10x (16 bp cell barcode and 10-12 bp UMI).

From the second sample, I see that _2 and _3 are identical, which probably means they are both R2.

In general, I would recommend reaching out to the authors of the data for clarification, as the organization is not entirely clear. Unfortunately, a lot of data in the SRA is not properly organized.

[linqy-immune]

Hi, thank you for your quick reply. Maybe such problem would be resulted from the SRA file.
I feel sad about my attempt. It is the first time I try to analyze single-cell data. I wonder whether it would be the fault of my fastq-dump code just like fastq-dump --split-3?
There are another 4 fastq files, which seem right organized, could you please help me have a look? Which files should be applied to mixcr analysis and what preset should I use? The extraction protocol is Single cell transcriptomic libraries were generated using the 5’v2 Gene expression and immune profiling kit (10x Genomics). Viable PBMCs were sorted into 2% FBS/PBS and cell counts were performed using a haemocytometer. Up to 40,000 cells were loaded into each lane of Chromium Next GEM Chip K Single Cell Kit (10x Genomics) to achieve a recovery cell number of approximately 20,000 cells. cDNA and TCR libraries were generated according to manufacturer’s instructions (10x Genomics). Generated libraries were sequenced on the NovaSeq S4 flow cell (Illumina) at Read 1 = 28, i7 index = 10, i5 index = 10 and Read 2: 90 cycles according to manufacturer’s instructions.

[mizraelson]

_3 is R1 and _4 is R2.

The command would be:

mixcr analyze 10x-sc-xcr-vdj \
     --species hsa \
     sample_R1.fastq.gz \
     sample_R2.fastq.gz \
     sample_result

[linqy-immune]

Got it! Thanks again for your advice.

[linqy-immune]

Hi Mark, @mizraelson
Long time no see. I am here again. I am a little confused whether the preset for such processing is generic-amplicon cause there is not specific preset for SuperScript VILO MasterMix?

[mizraelson]

Hi,
On the screenshot you shared it says Qiagen kit was used. You can use qiagen preset.

Sincerely,
Mark