If the two reads of a matepair map to the genome with the same orientations, this indicates an inversion.
If the mapped distance of a matepair is significantly different from the insert size of the matepair, then we may speculate that there is an insertion or a deletion (indel) between the pair of reads of the matepair.
If a particular matepair has k (concordant) mapped positions, then we count it as 1/ k of a matepair when computing the number of concordant matepairs mapped to any cluster.
To map each matepair to a unique cluster, we search for a valid configuration, ω, that has a one-to-one mapping between { X t } and { b(X t )}.
At each step of the algorithm, we find a matepair X t that we can move from its current cluster to another one, while increasing the objective function.
Perhaps more surprising than the similarity between our results and the previously described structural variants are the differences: we identify 3464 insertion, deletion, and inversion structural variants that do not overlap any structural variant in the DGV (Iafrate et al., 2004), of which 3032 have support from a uniquely mapped matepair.
Of these, 1374, 2279 and 185, respectively, were supported by at least a single uniquely mapped matepair.
We annotate a cluster as homozygous if it satisfies the following criteria: (1) no more than one uniquely mapped concordant matepair, (2) at least 4 discordant matepairs, and (3) at least a four-fold higher coverage by discordant matepairs than concordant matepairs.
There are two possible ways of mapping matepairs of genome A into REF. First, if we orient chromosomes p and q so that they have the same orientation, as shown in the figure, then the following conditions describe agreeing matepairs for a translocation: where a, b, c and d refer to the points in Figure 5.
Mapping a nation divided.
