How It Works

The scheme below shows how SMAP haplotype-sites is integrated with preprocessing, read mapping, locus delineation, and SNP calling. For GBS data, loci are positioned with SMAP delineate.

SMAP haplotype-sites reconstructs haplotypes based on SMAP positions and SNPs through read-backed haplotyping on a given set of BAM files.
SMAP haplotype-sites first creates sets of polymorphic positions per locus on the reference genome by intersecting locus regions (obtained with SMAP delineate) with a VCF file containing selected SNPs (obtained from any third-party SNP calling algorithm applied to the same set of BAM files).
Read-reference nucleotide pairs are retrieved by pysam ‘s get_aligned_pairs function, in which lower case nucleotides denote "different from the reference”.
In each BAM file, SMAP haplotype-sites then evaluates each read-reference alignment for the nucleotide aligned at the SMAP/SNP positions and scores as follows:

CALL TYPE	CLASSES
.	absence of read mapping
0	presence of the reference nucleotide
1	presence of an alternative nucleotide (any nucleotide different from the reference)
-	presence of a gap in the alignment

These calls are concatenated into a haplotype string of '.01-’s. For each discovered haplotype in the data, the total number of corresponding reads is counted per sample. Next, the haplotype counts of all samples are integrated into one master table, and expressed as relative haplotype frequency per locus per sample. Haplotypes with low frequency across all samples are removed to control for noise. The final table with haplotype frequencies per locus per sample is the end point for analysis of Pool-Seq data. Using the option --discrete_calls, SMAP haplotype-sites transforms the haplotype frequency table into discrete haplotype calls for individuals.

Three modes may be chosen for discrete haplotype calling in individuals:

CALL TYPE	CLASSES
dosage calls in diploids	0, 1, 2
dosage calls in triploids	0, 1, 2, 3
dosage calls in tetraploids	0, 1, 2, 3, 4
dosage calls in pentaploids	0, 1, 2, 3, 4, 5
dosage calls in pentaploids	0, 1, 2, 3, 4, 5, 6
dominant calls	0, 1

In the following sections, identification and quantification of haplotypes is illustrated on single-end GBS read data of a set of 8 diploid individuals at two partially overlapping loci. The content of the three example input files (BED, VCF, BAM) at this locus will be used to demonstrate the subsequent steps of SMAP haplotype-sites.