HTS: What is the correct way to access the sequence of a softclipped alignment?

[Juan Monroy-Nieto]

When an alignment is soft-clipped, extracting a location from the underlying sequence is shifted by the soft-clipped amount due to record.Pos assuming the masking while the extraction methods return the whole read. Both extraction methods for a sequence return the same right-shifted position.

As an example the CIGAR string would start with 3. unmapped leading bases as "?".
```txt
     T
XXXXXTXXXX //alignment
???XXWXXTX // record.Seq.Expand()
```

```go

//both will return W
r.Seq.Expand()[target - 1 - e.Pos]  //target is 1-based
r.Seq.At(target -1 e.Pos)
```

Alternative starting point using record.Start() does not work either because it matches record.Pos.

I have validated this findings using real data comparing Tablet GUI with all values. Currently I am generating the shift by parsing the cigar string. This is less than ideal due to relying on user (me) defined regexp.

The following is a real world example
```

# Tablet alignment info, seq soft-clipped
M02260:32:000000000-BLYVN:1:1101:18942:8135
From: 1,095 U1,095 to 1,383 U1,383
Length: 289 U289 (1 mismatch)
Cigar: 12S289M
Read Group: foo
Read direction is REVERSE

>M02260:32:000000000-BLYVN:1:1101:18942:8135
TTTACAACTTAGTATGGCAATATTTATATTCATTAAGAAAAGATAGAGCT
CCATTAGTGTTTTATTGGATTCCTTGGTTTGGTTCTGCAGCTTCATATGG
TCAACAACCTTATGAATTTTTTGAATCATGTCGTCAAAAGTATGGTGATG
TATTTTCATTTATGTTATTAGGGAAAATTATGACGGTTTATTTAGGTCCA
AAAGGTCATGAATTTGTTTTTAATGCTAAATTATCTGATGTTTCTGCTGA
AGATGCTTATAAACATTTAACTACTCCAGTTTTCGGTAA


# HTS/bam parsed
GCTCTTCCGATC //this 12bp portion is included but position still starts at 1094
TTTACAACTTAGTATGGCAATATTTATATTCATTAAGAAAAGATAGAGCT
CCATTAGTGTTTTATTGGATTCCTTGGTTTGGTTCTGCAGCTTCATATGG
TCAACAACCTTATGAATTTTT*T*GAATCATGTCGTCAAAAGTATGGTGATG //search target in between 2 stars
TATTTTCATTTATGTTATTAGGGAAAATTATGACGGTTTATTTAGGTCCA
AAAGGTCATGAATTTGTTTTTAATGCTAAATTATCTGATGTTTCTGCTGA
AGATGCTTATAAACATTTAACTACTCCAGTTTTCGGTAA

```

[Dan Kortschak]

If you can provide a minimal code reproducer, that would be really
helpful, but what I would suggest is that you get the offset by using
the Cigar.Lengths method, but only on the portion of the sam.Cigar that
includes the softclip.