Map reads to a reference

1. 01Map reads (FASTQ or FASTA) or contigs onto a reference sequence or a multi-record FASTA
2. 02Minimap2's presets for Illumina, Nanopore, PacBio HiFi and CLR, accurate long reads, Illumina Complete Long Reads and assemblies at three divergence levels
3. 03Inversions inside long reads found and written as their own records, and a divergence estimate per alignment
4. 04All-vs-all overlaps between long reads, and PAF output with or without base alignment
5. 05RNA-seq and cDNA aligned across introns, with every splice junction, its motif and its read support listed and selectable on the map
6. 06Paired-end mapping: mates paired record for record, proper pairs called within a fragment-length cap, and a uniquely placed mate lifting its partner in a repeat
7. 07SAM output with MAPQ, CIGAR, NM, alignment scores, supplementary splits and optional secondary records
8. 08Every minimap2 option adjustable under its own flag, defaulting to the chosen preset's values
9. 09A reading of the run: reads mapped, proper pairs, MAPQ 0 placements, repeats, splits, unmapped reads and reference coverage
10. 10Opens the alignment straight into the pileup against the same reference
11. 11The sample's consensus sequence called from the reads as samtools consensus does, as FASTA, FASTQ or pileup, opened as a new sequence

Reproduces minimap2's mapping (minimizer seeding, chaining, banded dual-affine alignment, inversion detection, MAPQ and pairing) under its short-read, RNA (spliced), long-read and assembly presets. Checked record for record against minimap2 2.31 itself on real human DNA and RNA-seq, RefSeq transcripts, E. coli, PacBio, mitochondrial and simulated reads. The panel runs on @ports/bam, @ports/map.

Everything runs client-side. Files are parsed in the browser and projects are saved on this machine, so an unpublished sequence is never uploaded.

What does MAPQ 0 mean?

The read aligns equally well in more than one place, so the position it was given is one of several and should not be trusted on its own. Reads inside repeats get MAPQ 0; a mate that maps uniquely can lift its partner above it.

Why did some reads not map?

Reads map only to what the reference contains. Reads from the host genome, from contamination, or from adapter dimers have nowhere to go, and a reference that is only part of a genome leaves every read from the rest of it unmapped.

What is a proper pair?

Both mates mapped to the same reference, facing each other, and close enough together (within the fragment-length setting) to have come from one DNA fragment.

What are supplementary alignments?

One read whose halves map to different places: the second piece is written as a supplementary record. It marks a junction the reference does not have, such as an insertion boundary, a rearrangement or a chimeric read.