Only small number of aligned reads are used to assemble in “generic ht single cell fragmented”

[bigcat1001]

Hi!
I have a strange question when using preset “generic ht single cell fragmented”.
The alignment looks fine, aligned reads: 2266062 (66.37%),but only 6.36% aligned reads aligned to CDR3.
And I can see in the assembly report that the number of reads used in the final clone is not very high, reads used in clones, percentage of total: 296775 (8.69%),which indicates that a large number of reads were not used in the assembly
How can I see where other reads have been aligned?

Analysis time: 1.71m
Total sequencing reads: 3414365
Successfully aligned reads: 2266062 (66.37%)
Coverage (percent of successfully aligned):
CDR3: 144201 (6.36%)
FR3_TO_FR4: 10312 (0.46%)
CDR2_TO_FR4: 8026 (0.35%)
FR2_TO_FR4: 175 (0.01%)
CDR1_TO_FR4: 39 (0%)
VDJRegion: 92 (0%)
Alignment failed: no hits (not TCR/IG?): 1134694 (33.23%)
Alignment failed after alignment-aided overlap: 6 (0%)
Alignment failed: low total score: 13603 (0.4%)
Overlapped: 2026366 (59.35%)
Overlapped and aligned: 1743232 (51.06%)
Overlapped and not aligned: 283134 (8.29%)
Alignment-aided overlaps, percent of overlapped and aligned: 44056 (2.53%)
No CDR3 parts alignments, percent of successfully aligned: 1912069 (84.38%)
Partial aligned reads, percent of successfully aligned: 209792 (9.26%)
V gene chimeras: 272732 (7.99%)
J gene chimeras: 38 (0%)
Paired-end alignment conflicts eliminated: 1713 (0.05%)
Realigned with forced non-floating bound: 2864110 (83.88%)
Realigned with forced non-floating right bound in left read: 38953 (1.14%)
Realigned with forced non-floating left bound in right read: 38953 (1.14%)
IGH chains: 2266062 (100%)
IGH non-functional: 1481 (0.07%)
Trimming report:
R1 reads trimmed left: 24 (0%)
R1 reads trimmed right: 147 (0%)
Average R1 nucleotides trimmed left: 3.7722973378651666E-4
Average R1 nucleotides trimmed right: 5.97182785085953E-4
R2 reads trimmed left: 636 (0.02%)
R2 reads trimmed right: 476 (0.01%)
Average R2 nucleotides trimmed left: 7.644173953282675E-4
Average R2 nucleotides trimmed right: 3.2041096953606306E-4
Tag parsing report:
Execution time: 0ns
Total reads: 3414365
Matched reads: 3414365 (100%)
Projection +R1 +R2: 3414365 (100%)
For variant 0:
For projection +R1 +R2:
R1:Left position: 7
R1:Right position: 150
CELL:Left position: 0
R2:Left position: 22
CELL:Right position: 22
Variants: 0
Cost: 0
R1 length: 143
CELL length: 22
R2 length: 128
======================================

Analysis time: 59.76s
Final clonotype count: 17661
Reads used in clonotypes, percent of total: 296775 (8.69%)
Average number of reads per clonotype: 16.8
Reads dropped due to the lack of a clone sequence, percent of total: 1729586 (50.66%)
Reads dropped due to a too short clonal sequence, percent of total: 6 (0%)
Reads dropped due to low quality, percent of total: 0 (0%)
Reads dropped due to failed mapping, percent of total: 73604 (2.16%)
Reads dropped with low quality clones, percent of total: 1797 (0.05%)
Aligned reads processed: 372596
Reads used in clonotypes before clustering, percent of total: 297189 (8.7%)
Number of reads used as a core, percent of used: 297189 (100%)
Mapped low quality reads, percent of used: 0 (0%)
Reads clustered in PCR error correction, percent of used: 0 (0%)
Reads pre-clustered due to the similar VJC-lists, percent of used: 0 (0%)
Clonotypes dropped as low quality: 219
Clonotypes eliminated by PCR error correction: 0
Clonotypes pre-clustered due to the similar VJC-lists: 0
Clones dropped in post filtering: 130 (0.73%)
Reads dropped in post filtering: 414.0 (0.01%)
Alignments filtered by tag prefix: 272 (0.01%)
IGH chains: 17661 (100%)
IGH non-functional: 173 (0.98%)
Pre-clone assembler report:
Number of input groups: 160412
Number of input groups with no assembling feature: 118784
Number of input alignments: 1853202
Number of alignments with assembling feature: 123616 (6.67%)
Number of output pre-clones: 51286
Number of pre-clonotypes per group:
0: + 26 (0.06%) = 26 (0.06%)
1: + 34480 (82.83%) = 34506 (82.89%)
2: + 5474 (13.15%) = 39980 (96.04%)
3: + 1118 (2.69%) = 41098 (98.73%)
4: + 281 (0.68%) = 41379 (99.4%)
5: + 114 (0.27%) = 41493 (99.68%)
6: + 135 (0.32%) = 41628 (100%)
Number of assembling feature sequences in groups with zero pre-clonotypes: 3641
Number of dropped pre-clones by tag suffix conflict: 0
Number of dropped alignments by tag suffix conflict: 0
Number of core alignments: 119092 (6.43%)
Discarded core alignments: 4524 (3.8%)
Empirically assigned alignments: 253253 (13.67%)
Empirical assignment conflicts: 1584 (0.09%)
Tag+VJ-gene empirically assigned alignments: 254837 (13.75%)
VJ-gene empirically assigned alignments: 0 (0%)
Tag empirically assigned alignments: 0 (0%)
Number of ambiguous groups: 7122
Number of ambiguous V-genes: 6431
Number of ambiguous J-genes: 4830
Number of ambiguous tag+V/J-gene combinations: 11261
Ignored non-productive alignments: 0 (0%)
Unassigned alignments: 320202 (17.28%)
======================================

[mizraelson]

Hi,
Since the data is fragmented (judging by the preset you have used) it is expected that only a small part of reads cover CDR3 region. Only 6.36% percent of 66% of total reads covers CDR3. The rest of the 66% cover other regions.
For this preset clones are initially assembled by CDR3 sequence, that is why assemble report shows a low number of reads used. Then assembleContigs (the next step in the pipeline) extends the initial CDR3 clones to cover the rest of the receptor sequence using the reads that were not used during the assembly.

[bigcat1001]

Thanks for you reply
AssembleContigs.report is below.But i didn't find how many reads was used.How can I determined the assemble quality?

Analysis time: 19.95s
Initial clonotype count: 17661
Final clonotype count: 17661 (100%)
Canceled assemblies: 0 (0%)
Number of premature termination assembly events, percent of number of initial clonotypes: 68.0 (0.39%)
Longest contig length: 752
Clustered variants: 0 (0%)
Reads in clustered variants: 0.0 (0%)
Reads in divided (newly created) clones: 0.0 (0%)
Clones with ambiguous letters in splitting region: 0 (0%)
Reads in clones with ambiguous letters in splitting region: 0.0 (0%)
Average number of ambiguous letters per clone with ambiguous letters in splitting region: NaN

[mizraelson]

what number do you get if you sum up the readCounts in the clonotype table?

[bigcat1001]

Thanks,I ignored this.Only 296775 reads were used for final clones which is about 10% of alignment reads(296775/2266062 ).
I will try to extract some reads to check what happened.
Based on

Coverage (percent of successfully aligned):
CDR3: 144201 (6.36%)
FR3_TO_FR4: 10312 (0.46%)
CDR2_TO_FR4: 8026 (0.35%)
FR2_TO_FR4: 175 (0.01%)
CDR1_TO_FR4: 39 (0%)
VDJRegion: 92 (0%)

I suppose that too many reads just cover 5' UTR? Do you have any suggestion?

[mizraelson]

This part doesn't say anything about 5'UTR. It shows that 6% cover CDR3, and then the longer the region the lower the percentage of reads cover it which is expected with fragmented data. What protocol did you use?

[bigcat1001]

The protocol I used similar to SMARTer3. Briefly,I prepared full length library using 5' race(similar to the TAKARA kit).Then,Tn5 with adaptors was added to fragment the full length library and a low number of third PCR cycles were used to amplify the fragment and sequcing using PE150.The final library structrue is just like single cell.

[bigcat1001]

The aim of this protocol is try to recover full length VDJ based on PE150 platform,not PE300.

[mizraelson]

But is it a single cell library? Or the barcode is a UMI?

[bigcat1001]

No,it is a bulk library.The barcode is a UMI. When i try the preset.I set UMIs as cell barcode.

[mizraelson]

Can you try the following: unzip the yaml file attached, put it into the directory you run MiXCR from (or to ~/.mixcr/presets/) and run the command:

mixcr analyze local:generic-shotgun-with-umi \
--species hsa \
--tag-pattern "^N{7}(R1:*)\^(UMI:N{22}(R2:*)"
input_R1.fastq.gz \
input_R2.fastq.gz \
result

Adjust species and pattern if needed (but use UMI not CELL barcode).

generic-shotgun-with-umi.yaml.zip