Read mapping is a straightforward application of the
UBLAST algorithm, though the default options may not give an optimal balance
between sensitivity and speed. Use the ublast command
with a UDB database. Assuming a typical scenario
where the reads have high identity with a reference genome, I recommend building
the database with -wordlength 14. Further speed improvements may be obtained by
reducing -dbaccel, I suggest testing values in the range 70 to 90. My testing
shows that these parameters give speeds that are typically around 50x faster
than bowtie with better sensitivity.