marginAlign is a probabilistic aligner that uses an input guide alignment as a starting point then realignes the
bases with a hidden Markov model (HMM). The HMM is trained using expectation maximization. The resulting alignments are
typically more accurate.
By default marginAlign chains the guide alignment so that each read has one maximal (by length) alignment to the
reference.
| Chrom | BAM |
|---|---|
| Chr7 | BAM |
| Chr20 | BAM |
| Chr22 | BAM |
These alignments were produced by learning the HMM from the guide (BWA) alignment then realigning, the chaining step was skipped and so each read may have multiple alignments
| Chrom | BAM |
|---|---|
| Chr7 | BAM |