This repository contains results from processing the SRR6750056 SPLiTSeq experiment with a hybrid Pheniqs/STARsolo approach. Command are assumed to be running on a machine with 16 cores.
The following are quoted from the STAR manual.
nNinBarcode: number of reads with more than 2 Ns in cell barcode (CB)
nUMIhomopolymer: number of reads with homopolymer in CB
nNoMatch: number of reads with CBs that do not match whitelist even with one mismatch
Following reads are discarded from Solo output.
Remaining reads are checked for overlap with features.
nUnmapped: number of reads unmapped to the genome
nNoFeature: number of reads that map to the genome but do not belong to a feature
nAmbigFeature: number of reads that belong to more than one feature
nAmbigFeatureMultimap: number of reads that belong to more than one feature and are also multimapping to the genome (this is a subset of the nAmbigFeature)
nTooMany: number of reads with ambiguous CB (i.e. CB matches whitelist with one mismatch but with posterior probability ¡0.95)
nNoExactMatch: number of reads with CB that matches a whitelist barcode with 1 mismatch, but this whitelist barcode does not get any other reads with exact matches of CB
Following reads are output in feature (e.g. gene) / cell count matrices.
nExactMatch: number of reads with CB that match the whitelist exactly
nMatch: total number of reads that match CB with 0 or 1 mismatches (this is superset of nExactMatch)
nCellBarcodes: number of distinct CBs detected
nUMIs: number of distinct UMIs detected
nNinBarcode+nUMIhomopolymer+nNoMatch+nTooMany+nNoExactMatch = number of reads with CBs that do not match whitelist.
nUnmapped+nAmbigFeature = number of reads without defined feature (gene)
nMatch = number of reads that are output as solo counts
The three categoties above summed together should be equal to the total number of reads.
This is a Basic STARsolo configuration for processing a SPLiTSeq experiment with the default strategy and produce the gene expression matrix.
STAR \
--runThreadN 16 \
--readFilesCommand bzcat \
--genomeDir /media/terminus/home/lg/split-seq-data/index/mm10_hg38 \
--soloFeatures Gene \
--readFilesPrefix /home/lg/split-seq-data/raw_fastq/ \
--readFilesIn SRR6750056_1.fastq.bz2 SRR6750056_2.fastq.bz2 \
--soloType CB_UMI_Complex \
--soloCBwhitelist /home/lg/code/moonwatcher/pheniqs-split-pool/raw/cb_whitelist \
/home/lg/code/moonwatcher/pheniqs-split-pool/raw/cb_whitelist \
/home/lg/code/moonwatcher/pheniqs-split-pool/raw/cb_whitelist \
--soloCBposition 0_10_0_17 0_48_0_55 0_86_0_93 \
--soloUMIposition 0_0_0_9 \
--soloCBmatchWLtype 1MM \
--outSAMtype BAM SortedByCoordinate \
--outSAMattributes CR CY UR UY GX GN CB UB sM sS sQ \cat usecase/solo/SRR6750056/A/Solo.out/Gene/Features.stats
nUnmapped 22655073
nNoFeature 116973794
nAmbigFeature 7457400
nAmbigFeatureMultimap 7431342
nTooMany 0
nNoExactMatch 1249
nExactMatch 3442881
nMatch 11316706
nMatchUnique 3859306
nCellBarcodes 73854
nUMIs 1338194cat solo/SRR6750056/A/Solo.out/Barcodes.stats
nNoAdapter 0
nNoUMI 0
nNoCB 0
nNinCB 64186
nNinUMI 108727
nUMIhomopolymer 3443228
nTooMany 0
nNoMatch 62544692
nMismatchesInMultCB 1575925
nExactMatch 133336154
nMismatchOneWL 17610668
nMismatchToMultWL 0This configuration for processing a SPLiTSeq experiment decodes the cellular barcodes with Pheniqs PAMLD and feeds the output directly into STAR for further processing over a standard stream without a temporary file. All 3 cellular barcodes are written to the CB/CY tags, and UMIs to RX/QX. Pheniqs drops reads that where any of the cellular barcodes fail to decode so STARsolo will see a reduced number of input reads but reads will have error corrected cellular barcodes so STAR can use an exact match strategy.
Configuration pheniqs-A uses PAMLD with 0.99 confidence threshold and 0.05 noise prior in all 3 cellular decoders and the default STARsolo strategy and produce the gene expression matrix. Configuration pheniqs-B is similar but uses a 0.95 confidence threshold for cellular decoding. Configuration pheniqs-C uses the same configuration as pheniqs-A to estimate the barcode and noise priors for all 3 decoders and uses the adjusted configuration for when feeding decoded reads to STARSolo. Configuration pheniqs-D uses the same configuration as pheniqs-B to estimate the barcode and noise priors.
This is a STARsolo/Pheniqs hybrid configuration for processing a SPLiTSeq experiment with Pheniqs decoding the 3 cellular barcodes using PAMLD with 0.99 confidence threshold and 0.05 noise prior and the default STARsolo strategy and produce the gene expression matrix.
STAR \
--runThreadN 16 \
--readFilesCommand "pheniqs mux --threads 12 --config splitseq_decode.json --sense-input --input" \
--genomeDir /media/terminus/home/lg/split-seq-data/index/mm10_hg38 \
--soloFeatures Gene \
--readFilesIn /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--soloType CB_UMI_Simple \
--soloCBwhitelist None \
--soloBarcodeMate 2 \
--soloCBlen 24 \
--soloUMIlen 10 \
--soloUMIstart 25 \
--soloInputSAMattrBarcodeSeq CB RX \
--soloInputSAMattrBarcodeQual CY QX \
--soloCBmatchWLtype Exact \
--readFilesType SAM SE \
--outSAMtype BAM SortedByCoordinate \
--readFilesSAMattrKeep RX QX CB CR CY XC \
--outSAMattributes GX GN CB UB sM sS sQ \cat pheniqs/SRR6750056/A/Solo.out/Gene/Features.stats
nUnmapped 20218934
nNoFeature 106847613
nAmbigFeature 6920453
nAmbigFeatureMultimap 6896585
nTooMany 0
nNoExactMatch 0
nExactMatch 3552197
nMatch 10472650
nMatchUnique 3552197
nCellBarcodes 69728
nUMIs 1265609This is a STARsolo/Pheniqs hybrid configuration for processing a SPLiTSeq experiment with Pheniqs decoding the 3 cellular barcodes using PAMLD with 0.95 confidence threshold and 0.05 noise prior and the default STARsolo strategy and produce the gene expression matrix.
STAR \
--runThreadN 16 \
--readFilesCommand "pheniqs mux --threads 12 --config splitseq_decode.json --sense-input --input" \
--genomeDir /media/terminus/home/lg/split-seq-data/index/mm10_hg38 \
--soloFeatures Gene \
--readFilesIn /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--soloType CB_UMI_Simple \
--soloCBwhitelist None \
--soloBarcodeMate 2 \
--soloCBlen 24 \
--soloUMIlen 10 \
--soloUMIstart 25 \
--soloInputSAMattrBarcodeSeq CB RX \
--soloInputSAMattrBarcodeQual CY QX \
--soloCBmatchWLtype Exact \
--readFilesType SAM SE \
--outSAMtype BAM SortedByCoordinate \
--readFilesSAMattrKeep RX QX CB CR CY XC \
--outSAMattributes GX GN CB UB sM sS sQ \cat pheniqs/SRR6750056/B/Solo.out/Gene/Features.stats
nUnmapped 20335194
nNoFeature 107246018
nAmbigFeature 6945574
nAmbigFeatureMultimap 6921625
nTooMany 0
nNoExactMatch 0
nExactMatch 3565063
nMatch 10510637
nMatchUnique 3565063
nCellBarcodes 70192
nUMIs 1270074This is a STARsolo/Pheniqs hybrid configuration for processing a SPLiTSeq experiment with Pheniqs decoding the 3 cellular barcodes using PAMLD and estimated priors proceeded by the default STARsolo strategy to produce the gene expression matrix.
An initial pheniqs run with discarded output ( --output /dev/null ) uses the same configuration as pheniqs-A to estimate the barcode and noise priors.
pheniqs mux \
--threads 16 \
--config splitseq_decode.json \
--sense-input \
--input /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--output /dev/null \
--prior splitseq_adjusted.jsonThe result of this run is a new configuration file, splitseq_adjusted.json that includes the imputed barcode and noise priors for all 3 cellular decoders. The bootstrapping configuration uses a 0.99 confidence threshold and 0.05 noise prior.
A subsequent run uses the new, prior adjusted, configuration file when executing pheniqs, overriding output to stdout (which would otherwise inherit the behavior from the initial run and redirect to /dev/null)
STAR \
--runThreadN 16 \
--readFilesCommand "pheniqs mux --threads 12 --config splitseq_adjusted.json --output /dev/stdout --sense-input --input" \
--genomeDir /media/terminus/home/lg/split-seq-data/index/mm10_hg38 \
--soloFeatures Gene \
--readFilesIn /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--soloType CB_UMI_Simple \
--soloCBwhitelist None \
--soloBarcodeMate 2 \
--soloCBlen 24 \
--soloUMIlen 10 \
--soloUMIstart 25 \
--soloInputSAMattrBarcodeSeq CB RX \
--soloInputSAMattrBarcodeQual CY QX \
--soloCBmatchWLtype Exact \
--readFilesType SAM SE \
--outSAMtype BAM SortedByCoordinate \
--readFilesSAMattrKeep RX QX CB CR CY XC \
--outSAMattributes GX GN CB UB sM sS sQ \cat pheniqs/SRR6750056/C/Solo.out/Gene/Features.stats
nUnmapped 19853238
nNoFeature 105006515
nAmbigFeature 6773231
nAmbigFeatureMultimap 6749822
nTooMany 0
nNoExactMatch 0
nExactMatch 3490017
nMatch 10263248
nMatchUnique 3490017
nCellBarcodes 68748
nUMIs 1251339This is a STARsolo/Pheniqs hybrid configuration for processing a SPLiTSeq experiment with Pheniqs decoding the 3 cellular barcodes using PAMLD and estimated priors proceeded by the default STARsolo strategy to produce the gene expression matrix.
An initial pheniqs run with discarded output ( --output /dev/null ) uses the same configuration as pheniqs-A to estimate the barcode and noise priors.
pheniqs mux \
--threads 16 \
--config splitseq_decode.json \
--sense-input \
--input /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--output /dev/null \
--prior splitseq_adjusted.jsonThe result of this run is a new configuration file, splitseq_adjusted.json that includes the imputed barcode and noise priors for all 3 cellular decoders. The bootstrapping configuration uses a 0.95 confidence threshold and 0.05 noise prior.
A subsequent run uses the new, prior adjusted, configuration file when executing pheniqs, overriding output to stdout (which would otherwise inherit the behavior from the initial run and redirect to /dev/null)
STAR \
--runThreadN 16 \
--readFilesCommand "pheniqs mux --threads 12 --config splitseq_adjusted.json --output /dev/stdout --sense-input --input" \
--genomeDir /media/terminus/home/lg/split-seq-data/index/mm10_hg38 \
--soloFeatures Gene \
--readFilesIn /home/lg/split-seq-data/raw_cram/SRR6750056.cram \
--soloType CB_UMI_Simple \
--soloCBwhitelist None \
--soloBarcodeMate 2 \
--soloCBlen 24 \
--soloUMIlen 10 \
--soloUMIstart 25 \
--soloInputSAMattrBarcodeSeq CB RX \
--soloInputSAMattrBarcodeQual CY QX \
--soloCBmatchWLtype Exact \
--readFilesType SAM SE \
--outSAMtype BAM SortedByCoordinate \
--readFilesSAMattrKeep RX QX CB CR CY XC \
--outSAMattributes GX GN CB UB sM sS sQ \cat pheniqs/SRR6750056/D/Solo.out/Gene/Features.stats
nUnmapped 20232480
nNoFeature 106929864
nAmbigFeature 6923081
nAmbigFeatureMultimap 6899201
nTooMany 0
nNoExactMatch 0
nExactMatch 3554858
nMatch 10477939
nMatchUnique 3554858
nCellBarcodes 69745
nUMIs 1266472cat pheniqs/SRR6750056/A/Solo.out/Gene/Summary.csv
Number of Reads,137582581
Reads With Valid Barcodes,1
Sequencing Saturation,0.643711
Q30 Bases in CB+UMI,-nan
Q30 Bases in RNA read,0.940886
Reads Mapped to Genome: Unique+Multiple,0.852994
Reads Mapped to Genome: Unique,0.729604
Reads Mapped to Gene: Unique+Multipe Gene,0.076119
Reads Mapped to Gene: Unique Gene,0.0258187
Estimated Number of Cells,4632
Unique Reads in Cells Mapped to Gene,3041312
Fraction of Unique Reads in Cells,0.856178
Mean Reads per Cell,656
Median Reads per Cell,525
UMIs in Cells,1076856
Mean UMI per Cell,232
Median UMI per Cell,186
Mean Gene per Cell,127
Median Gene per Cell,112
Total Gene Detected,14662
cat pheniqs/SRR6750056/B/Solo.out/Gene/Summary.csv
Number of Reads,138136950
Reads With Valid Barcodes,1
Sequencing Saturation,0.643744
Q30 Bases in CB+UMI,-nan
Q30 Bases in RNA read,0.940398
Reads Mapped to Genome: Unique+Multiple,0.85274
Reads Mapped to Genome: Unique,0.729369
Reads Mapped to Gene: Unique+Multipe Gene,0.0760885
Reads Mapped to Gene: Unique Gene,0.0258082
Estimated Number of Cells,4638
Unique Reads in Cells Mapped to Gene,3053375
Fraction of Unique Reads in Cells,0.856472
Mean Reads per Cell,658
Median Reads per Cell,527
UMIs in Cells,1080803
Mean UMI per Cell,233
Median UMI per Cell,187
Mean Gene per Cell,127
Median Gene per Cell,112
Total Gene Detected,14695
cat pheniqs/SRR6750056/C/Solo.out/Gene/Summary.csv
Number of Reads,135162095
Reads With Valid Barcodes,1
Sequencing Saturation,0.641452
Q30 Bases in CB+UMI,-nan
Q30 Bases in RNA read,0.94326
Reads Mapped to Genome: Unique+Multiple,0.853072
Reads Mapped to Genome: Unique,0.729967
Reads Mapped to Gene: Unique+Multipe Gene,0.0759329
Reads Mapped to Gene: Unique Gene,0.025821
Estimated Number of Cells,4626
Unique Reads in Cells Mapped to Gene,2988451
Fraction of Unique Reads in Cells,0.856286
Mean Reads per Cell,646
Median Reads per Cell,516
UMIs in Cells,1064884
Mean UMI per Cell,230
Median UMI per Cell,184
Mean Gene per Cell,126
Median Gene per Cell,111
Total Gene Detected,14597
cat pheniqs/SRR6750056/D/Solo.out/Gene/Summary.csv
Number of Reads,137683698
Reads With Valid Barcodes,1
Sequencing Saturation,0.643735
Q30 Bases in CB+UMI,-nan
Q30 Bases in RNA read,0.940733
Reads Mapped to Genome: Unique+Multiple,0.853004
Reads Mapped to Genome: Unique,0.729618
Reads Mapped to Gene: Unique+Multipe Gene,0.0761015
Reads Mapped to Gene: Unique Gene,0.025819
Estimated Number of Cells,4633
Unique Reads in Cells Mapped to Gene,3044102
Fraction of Unique Reads in Cells,0.856322
Mean Reads per Cell,657
Median Reads per Cell,525
UMIs in Cells,1077748
Mean UMI per Cell,232
Median UMI per Cell,186
Mean Gene per Cell,127
Median Gene per Cell,112
Total Gene Detected,14671
cat solo/SRR6750056/A/Solo.out/Gene/Summary.csv
Number of Reads,218683580
Reads With Valid Barcodes,0.690246
Sequencing Saturation,0.653255
Q30 Bases in CB+UMI,0.860847
Q30 Bases in RNA read,0.937268
Reads Mapped to Genome: Unique+Multiple,0.764201
Reads Mapped to Genome: Unique,0.645298
Reads Mapped to Gene: Unique+Multipe Gene,0.0517492
Reads Mapped to Gene: Unique Gene,0.0176479
Estimated Number of Cells,4697
Unique Reads in Cells Mapped to Gene,3316833
Fraction of Unique Reads in Cells,0.859438
Mean Reads per Cell,706
Median Reads per Cell,567
UMIs in Cells,1142714
Mean UMI per Cell,243
Median UMI per Cell,195
Mean Gene per Cell,132
Median Gene per Cell,117
Total Gene Detected,14968