Updates from review.

- Fixed typos / grammar.
- Require strand to be indicative of the original sequence orientation
  and not the current one listed in the file.
- Added example text for multiple choices of modification at a single
  site.
- Added space as a separator between MP entries.
- Adjusted the regexp to permit empty numeric series (representing no
  modification present).

SAMtags.tex

@@ -457,7 +457,7 @@ \subsubsection{Color space}
 
 \subsection{Base modifications}
 
-Base basications, including base methylation, are represented as a
+Base modifications, including base methylation, are represented as a
 series of edits from the primary unmodified sequence stored in the
 main SAM {\sf SEQ} field.  If the modifed base has no natural
 unmodified form then this should be stored as ``N''.
@@ -468,29 +468,32 @@ \subsection{Base modifications}
 this modification being correct, rather than the base being unmodified.
 
 \begin{description}
-\item[MM:Z:\tagregex{([ACGTN][-+][a-z](,[0-9]+)+;)*}]
+\item[MM:Z:\tagregex{([ACGTN][-+][a-z](,[0-9]+)*;)*}]
 \hfill\\
 The first character is the unmodified base as seen in the {\sf SEQ}
 field, one of {\tt A}, {\tt C}, {\tt G}, {\tt T} or {\tt N}, with
 the exception that {\tt N} is used to match any base rather than
 strictly {\tt N}.  This is followed by either plus or minus
 indicating the strand the modification was observed on (relative to
-the recorded strand of {\sf SEQ} with plus meaning same
-orientation), and a base modification symbol.  This is then followed
-by a comma separated list of how many unmodified seq bases of the
+the original sequenced strand of {\sf SEQ} with plus meaning same
+orientation\footnote{Hence a tool that may reverse complement sequences does not need to understand how to manipulate the {\tt MM} and {\tt MP} tags.}), and a base modification symbol.
+This is then followed by a comma separated list of how many unmodified seq bases of the
 stated base type to skip, stored as a delta to the last and starting
 with 0 as the first (or next) base.  Hence this number series is
 comparable to the numbers in an {\tt MD} tag.
 
 For example {\tt C+m,5,12,0;} tells us there are three 5-Methylcytosine
 bases in the original {\sf SEQ}.  The first 5 {\tt C} bases are
-unmodified and the 6th is modified, as are the 19th (12 inbetween the
+unmodified and the 6th is modified, as are the 19th (12 between the
 6th and 19th) and 20th.  Similarly {\tt G-m,14;} indicates the 15th
 {\tt G} is a 5-Methylcytosine on the opposite strand.
 
 This permits modifications to be listed on either strand with the rare
 potential for both strands to have a modification at the same site.
 
+Note it is permitted for the coordinate list to be empty (for example {\tt MM:Z:C+m;}), which may be used as an explicit indicator that this base modification is not present.
+It is not permitted for coordinates to be beyond the length of the sequence.
+
 If the modification is not one of the standard common types (listed
 below) it can be specified as a numeric ChEBI code.  For example
 {\tt C+76792,57;} is the same as {\tt C+h,57;}.
@@ -536,22 +539,25 @@ \subsection{Base modifications}
 
 \item[MP:Z:\tagvalue{qualities}]
 \hfill\\
-The {\tt MP} tag if present lists the Phred qualities of each
-modification listed in the {\tt MM} tag.  The length should match the
-number of position deltas from {\tt MM}.  The qualities are encoded in
-the same manner as the primary {\sf QUAL} field; one byte per quality
-with ASCII value Phred score + 33.  No separators should be present.
+The optional {\tt MP} tag lists the Phred qualities of each modification listed in the {\tt MM} tag.
+The qualities are encoded in the same manner as the primary {\sf QUAL} field; one byte per quality with ASCII value Phred score + 33.
+A space character (`{\tt \textvisiblespace}') should be used as a separator between concatenated quality strings when multiple modification lists are present in the {\tt MM} tag.
+The length should match the number of position deltas from {\tt MM} plus 1 per space character required.
 
 For example {\tt MM:Z:C+m,5,12,3;C+h,57;} may have an associated
-quality tag of {\tt MP:Z:5EB/}.
+quality tag of {\tt MP:Z:5EB /}.
 
 Quality values for ambiguity codes give the likelihood that the
 modification is one of the possible codes compatible with that
-ambiguity code.  For example {\tt MM:Z:C+C,10 MP:Z:+} indicates a C
+ambiguity code.  For example {\tt MM:Z:C+C,10; MP:Z:+} indicates a C
 call with an unspecified modification and the phred score of 10 (ASCII
 value {\tt +}).  This corresponds to a 90\% chance of the base being
 modified.
 
+To represent several possible modifications at the same site the {\tt MP} tag can be used to indicate the probabilities of each possibility.
+The values used should be absolute probabilities, not relative between the alternatives.
+For example, a C base that has 90\% chance of being modified with 5mC being three times more likely than 5hmC will encode 5mC with 67.5\% probability ($0.9 * 0.75$)and 5hmC with 22.5\% probability ($0.9 * 0.25$).
+This could be represented with {\tt MM:Z:C+m,10;C+h,10; MP:Z:" \&}.
 
 \end{description}