Markup | Genome Research

Table 1.

Steps in the Atlas Assembly System

Stage	Program	Action	Comment
1. Data preparation	Data quality checks	Contamination (reads from other organisms) and mislabeled reads (ie, from another BAC) are identified and corrected if possible.
	trim-reads	Remove low-quality bases, so that only highest-quality sequence is used for finding overlaps between reads.	Trimmed reads only used in finding overlaps; full sequences used to assemble the consensus sequence.
2. Analyze sequence Redundancy	k-mer-counter	Build table of the frequency of oligonucleotides (k-mers).	Only WGS reads used to give most complete and random sampling of the genome.
3. Compute read overlap graph	Overlapper	Identify candidate overlaps based on shared rare k-mers.	End-to-end criterion scores alignment on entire overlapping portion of reads.
		Stringently evaluate overlaps by banded alignment.
		Save overlap graph with stringency annotations.
4. eBAC assembly		Coassemble WGS and skim reads that have been assigned to the same BAC to produce eBACs.
	binner	Choose WGS reads with best overlaps to skim reads in a BAC; add read pair mates.
	Phrap	Assemble WGS and skim reads.
	split-scaffold	Split misjoined contigs.
	split-scaffold	Build scaffolds with read pairs.
5. Build bactigs		Find overlapping eBACs based on shared reads and more.
	BLASTZ	Confirm overlap by aligning eBACs.
		Compare bactigs to other maps for verification.
6. Assembly of bactigs	rolling-phrap	Assemble reads in bactigs.
	Phrap	Assemble contigs in bactigs.
	split-scaffold	Split misjoins, build scaffolds.
7. Build superbactigs		Link bactigs by read pairs and BAC skim read distribution.
8. Build ultrabactigs and map to chromosomes		Link superbactigs by map and synteny data.