danbing-tk v1.3.2

Major changes:

• Automated bias correction using danbing-tk-pred

Resources:

• ikmer.meta required by danbing-tk-pred
• ikmer.meta.txt human readable version of ikmer.meta.txt with format documented in Wiki
• Example trkmers.meta.txt required by danbing-tk-pred

Next release (v1.3.3):

• Automated dosage computation for motifs and TR loci

danbing-tk v1.3.1 (manuscript)

This version is associated with the manuscript: "The motif composition of variable-number tandem repeats impacts gene expression"

Major changes:

• Updated preferred usage of danbing-tk by turning on kmer filter: -kf 4 1
• Reduces *.tr.kmers output size by saving only counts, and uses index file to reconstruct locus name and kmer names

Resources in Assets:

• tr.good.bed: VNTR set for building RPGG

Additional resource on Zenodo:

• VNTR statistics and annotations on 35 HGSVC assemblies
• RPGG built from the annotations
• GTEx gene-level eVNTR discoveries
• GTEx gene-level eMotif discoveries
• GTEx fine-mapping results using susieR
• Bias matrices for HGSVC, HPRC, GTEx, and Geuvadis samples used in bias correction
• GTEx bias-corrected kmer dosage table
• Geuvadis bias-corrected kmer dosage table

Additional analysis scripts for bias correction, eQTL mapping, and fine-mapping are available in this repo.

danbing-tk v1.3

Improvements:

• Significantly improve the time/mem usage of danbing-tk
  • benchmark setting
    • 31x HG00731 SRS sample from 1000 Genomes Project
    • two-consortium RPGG, 81045 loci
    • 16 cores xeon-2665, avx
    • samtools fasta -@2 -n $bam | danbing-tk -a -kf 4 1 -gc 80 -k 21 -qs pan -fa /dev/stdin -o $out -p 16 -cth 45 | gzip >$aln
  • Sample was genotyped in ~43 min using 31.4 Gb mem
  • 24x speedup, 37% reduction in mem usage
  • Output file size: 1.3 Gb
• danbing-tk now takes binary graph/index as input
  • ktools serialze was added to convert *kmers to *.graph.umap *.kmerDBi.umap and *.kmerDBi.vv
• bam2pe is now merged with danbing-tk
  • use -fa option for non-interleaved fasta e.g. samtools fasta -@2 -n $bam
  • use -fai for interleaved fasta

Resources

• New RPGG and VNTR coordinates on 35 HGSVC genomes are available at Zenodo

danbing-tk v1.2

Improvements:

• Improved indel handling in graph threading.
• Improved the memory scalability of multiple-boundary-alignment.

Resources:

• New RPGG and VNTR coordinates on 35 HGSVC genomes are available at Zenodo

manuscript-1

Latest version of code and resources that associate with the manuscript "Profiling variable-number tandem repeat variation across populations using repeat-pangenome graphs". Released for creating DOI with Zenodo.

danbing-tk v1.1

Improvements:

• Faster danbing-tk aliign: 2.6x speedup on HG00096 when genotying 32,138 loci
• More flexible use of danbing-tk build: generating RPGG without SRS data by skipping graph pruning
• More informative aln-r2: fixed zero r2 when no variation in assembly kmer count by adding a dummy point at (0,0)

danbing-tk v1.0

Improvements:

• Improved length estimation accuracy using multi-boundary expansion, due to more accurate orthology mapping of VNTRs across haplotypes.
• More stringent QC on VNTR size, number of supporting haplotypes, consistency of liftover coordinates, etc.
• Slightly expand VNTR set from 29,111 to 32,138 loci.
• Added more user-friendly length estimation script.
• Added option for alignment output by using -a with danbing-tk align
• DOI created using Zenodo

Additional resources:

• Repeat-pangenome graph encoded as pan.tr.kmers, pan.ntr.kmers and pan.graph.kmers in RPGG.tar.gz
• 84,411 raw VNTR coordinates tr.84411.bed
• 32,138 raw VNTR coordinates (high-confidence genotypable set) tr.good.bed
• 397 non-VNTR regions ctrl.bed
• Locus-specific biases of VNTR and non-VNTR regions LSB.tsv
• Summary of eGene discoveries Alltissue.egenes.tsv
• Comprehensive VNTR statistics vntr.statistics.tsv vntr.statistics.README
• 13 PacBio CLR assemblies (26 haplotypes) *.h?.fasta.gz
• 32,138 boundary-expanded VNTR coordinates in the 26 haplotypes pan.tr.mbe.no_CCS.bed and pan.tr.mbe.no_CCS.README
• 73,582 boundary-expanded VNTR coordinates pan.tr.73582.mbe.no_CCS.bed

Example analyses:

• QC of multi-boundary expansion 202011.MultiBoundaryExpansion.QC.ipynb
• Measuring length prediction accuracy 202012.Acc.pan.ipynb
• Contrasting the most informative kmer between populations 202012.mikmer.ipynb
• eQTL mapping 202012.eQTL.32138.ipynb
• Sample QC on locus-specific bias LSB_analysis.ipynb
• Heritability analysis of SNP v.s. SNP+VNTR models 202011.sg.joint.ipynb
• Miscellaneous analyses in the original manuscript 202012.revision.supp.ipynb

v0.0

Version 0 of genotypable VNTRs, RPGG and precomputed LSB are out! These files should be the same as the ones used for the analysis in the original paper.