Question about output content

[BOBOL513]

Hi , I'm new to abyss trying to use short-read genome data to build genome assembly. A test was run with one sample from HGSVC3 (about 40x), two illumina short reads fastq.gz files as input. And the question is how to explain that the output contigs.fasta only has about 3G base pairs? I thought it would produce a complete diploid genome assembly.

The command was something like :

abyss-pe name=***** k=96 B=100G in='/absolute_path/sample_1.fastq.gz /absolute_path/sample_2.fastq.gz' -C /absolute_path/abyss_workflow

Thanks for developing this tool and answering my questions

[jwcodee]

Thanks for using ABySS. I am assuming your sample is from a human individual. If so, 3Gbp would be the correct size of the genome that ABySS assembles. ABySS only assembles a pseudo haploid of your genomic data, and not a diploid genome.

[BOBOL513]

Thanks for using ABySS. I am assuming your sample is from a human individual. If so, 3Gbp would be the correct size of the genome that ABySS assembles. ABySS only assembles a pseudo haploid of your genomic data, and not a diploid genome.

I see, also I want to konw how ABySS represent a heterozygous loci in the reported haploid assembly? Random selection of one allele should not make much difference for SNP and INDEL, but may generate very different assembly for structural variations.

[warrenlr]

Assuming a diploid genome, for large SVs, it depends on paired end read support. Typically, one allele representation will be favoured over the other if it is strongly supported by read pairs (or there is no path for the other allele in the graph). When that is not the case, the assembly may break at those positions.