Algorithm 1.

2SNV Algorithm

Procedure 1: constructing the consensus haplotype for all reads:

Initialize the set of all clusters with a single cluster with all reads

For each position i find allele of highest frequency ai

Procedure 2: partitioning reads into simple clusters

while not all clusters are simple do

for each non-simple cluster do

if no pair SNVs is linked according to (2–4) then

Regard C as a simple cluster

else

Find a pair of linked SNVs I2 and J2 minimizing (3)

Find the set C1 of all reads with the 2-haplotype

Find the consensus

,

Procedure 3: estimating frequencies of the consensuses of simple clusters

Run kGEM algorithm for the set of haplotypes .