Enhancing somatic variant calling and execution speed

[sloth-eat-pudding]

Hello,

I am working with HCC1395 data, analyzing tumor samples at 75x coverage and normal samples at 45x coverage. I utilized Clair3 to process the normal.bam file, generating a normal.vcf. This file was then employed for phasing and haplotagging the tumor.bam, followed by using a somatic mutation caller. The results showed a notable decrease in false positives.

phase and haplotag	Precision	Recall	F1-score	TP	FP	FN
ClairS germline.vcf	67.12%	77.64%	72.00%	30626	15001	8821
Clair3 normal.vcf	72.50%	77.46%	74.90%	30556	11593	8891

In an instance where false positives were converted to true negatives, it was observed that the mutations in the normal sample were heterozygous, whereas in the tumor sample, they were homozygous. This suggests a loss of heterozygosity (LOH) event, making the strategy of phasing and tagging most reads into the same haplotype seem correct. Have you considered this method?

Moreover, I noted in literature that the primary reason for choosing Longphase for phasing is its speed. We still have a speed advantage in haplotagging. ClairS employs parallel acceleration at the chromosome level and we can introduce a feature to specify a range. Could this reduce the training costs for you? I also conducted a haplotag test, and the results do not seem to show any significant differences.

haplotag	Precision	Recall	F1-score	TP	FP	FN
whatshap v1.7	67.12%	77.64%	72.00%	30626	15001	8821
longphase v1.3	67.27%	77.62%	72.07%	30617	14897	8830

[aquaskyline]

Hi longphase team. Thanks for asking. We spoke in more detail via email. ClairS is ready to make use of additional HP taggings in addition to the current HP1 and HP2. Basically, there is no limit to the number of HP categories ClairS can take. For parallelization, supporting range processing sounds good, ClairS will most likely use it in a per chromosome fashion.

[sloth-eat-pudding]

We have released version 1.7. Haplotag now includes the --region feature.

The complete list of haplotag parameters

Usage:  haplotag [OPTION] ... READSFILE
      --help                          display this help and exit.

require arguments:
      -s, --snp-file=NAME             input SNP vcf file.
      -b, --bam-file=NAME             input bam file.
      -r, --reference=NAME            reference fasta.
optional arguments:
      --tagSupplementary              tag supplementary alignment. default:false
      --sv-file=NAME                  input phased SV vcf file.
      --mod-file=NAME                 input a modified VCF file (produced by longphase modcall and processed by longphase phase).
      -q, --qualityThreshold=Num      not tag alignment if the mapping quality less than threshold. default:1
      -p, --percentageThreshold=Num   the alignment will be tagged according to the haplotype corresponding to most alleles.
                                      if the alignment has no obvious corresponding haplotype, it will not be tagged. default:0.6
      -t, --threads=Num               number of thread. default:1
      -o, --out-prefix=NAME           prefix of phasing result. default:result
      --region=REGION                 tagging include only reads/variants overlapping those regions. default:(all regions)
      --log                           an additional log file records the result of each read. default:false

[zhengzhenxian]

@sloth-eat-pudding Glad to have the new release of LongPhase, our team will test it and get back to you.

ZX

[sloth-eat-pudding]

In a https://github.com/HKU-BAL/ClairS/issues/18#issuecomment-1893671596 issue, it was mentioned that a "longer phaseset and an improved haplotagging ratio" were needed. Therefore, I attempted to incorporate indels for phasing and haplotagging.

Explanation of data sources

• snp.vcf : uses the confident heterozygous germline variants from ClairS.
• indle.vcf : generated by Clair3 calling. normal & tumor: used only if the chromosome, position, and genotype are identical. normal all : uses all indels from the normal.

snp.vcf	indel.vcf	phase	haplotag	Precision	Recall	F1-score	TP	FP	FN
ClairS	-	longphase v1.3	whatshap v1.7	67.12%	77.64%	72.00%	30626	15001	8821
ClairS	-	longphase v1.3	longphase v1.3	67.27%	77.62%	72.07%	30617	14897	8830
ClairS	-	longphase v1.7	whatshap v1.7	67.27%	77.62%	72.07%	30619	14899	8828
ClairS	indel (normal & tumor)	longphase v1.7-indel	whatshap v1.7	67.44%	77.57%	72.15%	30599	14770	8848
ClairS	indel (normal & tumor)	longphase v1.7-indel	longphase v1.3(no tag indel)	67.48%	77.57%	72.18%	30601	14745	8846
ClairS	indel (normal & tumor)	longphase v1.7-indel	longphase v1.7(tag indel)	67.75%	77.52%	72.31%	30578	14553	8869
ClairS	indel (normal all)	longphase v1.7-indel	longphase v1.7(tag indel)	68.80%	77.44%	72.87%	30548	13853	8899

Would you be interested in trying to incorporate indels as well?

[aquaskyline]

Yes, doing that in the next version.