Comparing Minimizers and MEMS as anchors in chaining

Chaining algorithm:

• modified ChainX, algbio-ChainX
• original ChainX, ChainX

Anchors:

• asymmetric MEM
• symmetric MEM, r-index, br-index-mem
• MUM
• Minimizers

Existing implementation of chaining + different anchors bdbwt-mem-chaining

Snakemake file is run with command

snakemake --cores 1

TODO:

• run on e coli sequence
• generate reads

  simlord --read-reference data/test/ecoli.fasta -n 10000 --no-sam data/test/ecoli_reads https://academic.oup.com/bioinformatics/article/32/17/2704/2450740

• confirm that anchors are matches

  mostly matches what is the fastest method to compute the anchors? > compare: 1. MUMMER, mem and mum, 2. bdbwt, extended minimizer, MEMs 3. r-index, MEM, MUM? 4. minimap2 minimizers, extended minimizers with minimap2 minimizers bdbwt is horrible computing mems when compared to chainX:s mummer, with e-coli.

• compute MUMS

  use the ChainX mummer implementation or implement ur own

• k mer size dependent on the length of the sequence as in the paper. 22 for e coli 31 for huma
• report number of anchors and the sum of anchor bases.
• compute the Jaccard index and draw siome figures

Stats to report:

• average length of reads
• average number of anchor per chain
• average length of anchor bases
• average number of anchors per read
• anchor coverage of read
• chain coverage of read
• Jaccard index (the shared bases between the chain and the read)/union of the chain and the read