import java.util.ArrayList;
import java.util.List;
+import jalview.bin.Cache;
+import jalview.bin.Console;
import jalview.datamodel.SequenceI;
/**
{
private static final int EIGHTY_FIVE = 85;
- private static final int NINETY_NINE = 99;
+ private static final int NUCLEOTIDE_COUNT_PERCENT;
- private static final int TO_UPPER_CASE = 'a' - 'A';
+ private static final int NUCLEOTIDE_COUNT_LONG_SEQUENCE_AMBIGUITY_PERCENT;
+
+ private static final int NUCLEOTIDE_COUNT_SHORT_SEQUENCE;
+
+ private static final int NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE;
+
+ private static final boolean NUCLEOTIDE_AMBIGUITY_DETECTION;
public static final char GAP_SPACE = ' ';
new char[]
{ GAP_SPACE, GAP_DOT, GAP_DASH });
+ static
+ {
+ // these options read only at start of session
+ NUCLEOTIDE_COUNT_PERCENT = Cache.getDefault("NUCLEOTIDE_COUNT_PERCENT",
+ 55);
+ NUCLEOTIDE_COUNT_LONG_SEQUENCE_AMBIGUITY_PERCENT = Cache.getDefault(
+ "NUCLEOTIDE_COUNT_LONG_SEQUENCE_AMBIGUITY_PERCENT", 95);
+ NUCLEOTIDE_COUNT_SHORT_SEQUENCE = Cache
+ .getDefault("NUCLEOTIDE_COUNT_SHORT", 100);
+ NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE = Cache
+ .getDefault("NUCLEOTIDE_COUNT_VERY_SHORT", 4);
+ NUCLEOTIDE_AMBIGUITY_DETECTION = Cache
+ .getDefault("NUCLEOTIDE_AMBIGUITY_DETECTION", true);
+ }
+
/**
* DOCUMENT ME!
*
jlen--;
}
- int count = 0;
int match = 0;
float pid = -1;
{
match++;
}
-
- count++;
}
pid = (float) match / (float) ilen * 100;
{
match++;
}
-
- count++;
}
pid = (float) match / (float) jlen * 100;
*/
public static final boolean isNucleotide(SequenceI seq)
{
- if (seq == null)
+ if (seq == null || seq.getLength() == 0)
{
return false;
}
- long ntCount = 0;
- long aaCount = 0;
- long nCount = 0;
- long ntaCount = 0;
+ long ntCount = 0; // nucleotide symbol count (does not include ntaCount)
+ long aaCount = 0; // non-nucleotide, non-gap symbol count (includes nCount
+ // and ntaCount)
+ long nCount = 0; // "Unknown" (N) symbol count
+ long xCount = 0; // Also used as "Unknown" (X) symbol count
+ long ntaCount = 0; // nucleotide ambiguity symbol count
int len = seq.getLength();
for (int i = 0; i < len; i++)
}
else
{
+ if (isX(c))
+ {
+ xCount++;
+ }
if (isNucleotideAmbiguity(c))
{
ntaCount++;
}
}
}
- /*
- * Check for nucleotide count > 85% of total count (in a form that evades
- * int / float conversion or divide by zero).
- */
- if ((ntCount + nCount) * 100 > EIGHTY_FIVE * (ntCount + aaCount))
+ long allCount = ntCount + aaCount;
+
+ if (NUCLEOTIDE_AMBIGUITY_DETECTION)
{
- return ntCount > 0; // all N is considered protein. Could use a threshold
- // here too
+ Console.debug("Performing new nucleotide detection routine");
+ if (allCount > NUCLEOTIDE_COUNT_SHORT_SEQUENCE)
+ {
+ // a long sequence.
+ // check for at least 55% nucleotide, and nucleotide and ambiguity codes
+ // (including N) must make up 95%
+ return ntCount * 100 > NUCLEOTIDE_COUNT_PERCENT * allCount
+ && 100 * (ntCount + nCount
+ + ntaCount) > NUCLEOTIDE_COUNT_LONG_SEQUENCE_AMBIGUITY_PERCENT
+ * allCount;
+ }
+ else if (allCount > NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE)
+ {
+ // a short sequence.
+ // check if a short sequence is at least 55% nucleotide and the rest of
+ // the symbols are all X or all N
+ if (ntCount * 100 > NUCLEOTIDE_COUNT_PERCENT * allCount
+ && (nCount == aaCount || xCount == aaCount))
+ {
+ return true;
+ }
+
+ // a short sequence.
+ // check for at least x% nucleotide and all the rest nucleotide
+ // ambiguity codes (including N), where x slides from 75% for sequences
+ // of length 4 (i.e. only one non-nucleotide) to 55% for sequences of
+ // length 100
+ return myShortSequenceNucleotideProportionCount(ntCount, allCount)
+ && nCount + ntaCount == aaCount;
+ }
+ else
+ {
+ // a very short sequence. (<4)
+ // all bases must be nucleotide
+ return ntCount > 0 && ntCount == allCount;
+ }
}
else
{
- // check for very large proportion of nucleotide and all ambiguity codes
- if ((ntCount + nCount + ntaCount) * 100 >= NINETY_NINE
- * (ntCount + aaCount))
+ Console.debug("Performing old nucleotide detection routine");
+ /*
+ * Check for nucleotide count > 85% of total count (in a form that evades
+ * int / float conversion or divide by zero).
+ */
+ if ((ntCount + nCount) * 100 > EIGHTY_FIVE * allCount)
{
- return ntCount > 0;
+ return ntCount > 0; // all N is considered protein. Could use a
+ // threshold here too
}
- return false;
}
+ return false;
+ }
+
+ protected static boolean myShortSequenceNucleotideProportionCount(
+ long ntCount, long allCount)
+ {
+ /**
+ * this method is valid only for NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE <=
+ * allCount <= NUCLEOTIDE_COUNT_SHORT_SEQUENCE
+ */
+ // the following is a simplified integer version of:
+ //
+ // a := allCount # the number of bases in the sequence
+ // n : = ntCount # the number of definite nucleotide bases
+ // vs := NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE
+ // s := NUCLEOTIDE_COUNT_SHORT_SEQUENCE
+ // lp := NUCLEOTIDE_COUNT_LOWER_PERCENT
+ // vsp := 1 - (1/a) # this is the proportion of required definite
+ // nucleotides
+ // # in a VERY_SHORT Sequence (4 bases).
+ // # This should be equivalent to all but one base in the sequence.
+ // p := (a - vs)/(s - vs) # proportion of the way between
+ // # VERY_SHORT and SHORT thresholds.
+ // tp := vsp + p * (lp/100 - vsp) # the proportion of definite nucleotides
+ // # required for this length of sequence.
+ // minNt := tp * a # the minimum number of definite nucleotide bases
+ // # required for this length of sequence.
+ //
+ // We are then essentially returning:
+ // # ntCount >= 55% of allCount and the rest are all nucleotide ambiguity:
+ // ntCount >= tp * allCount && nCount + ntaCount == aaCount
+ // but without going into float/double land
+ long LHS = 100 * allCount
+ * (NUCLEOTIDE_COUNT_SHORT_SEQUENCE
+ - NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE)
+ * (ntCount - allCount + 1);
+ long RHS = allCount * (allCount - NUCLEOTIDE_COUNT_VERY_SHORT_SEQUENCE)
+ * (allCount * NUCLEOTIDE_COUNT_PERCENT - 100 * allCount + 100);
+ return LHS >= RHS;
}
/**
return isNucleotide(c, false);
}
- public static boolean isNucleotide(char c, boolean countAmbiguity)
+ /**
+ * includeAmbiguity = true also includes Nucleotide Ambiguity codes
+ */
+ public static boolean isNucleotide(char c, boolean includeAmbiguity)
{
char C = Character.toUpperCase(c);
switch (C)
case 'U':
return true;
}
- if (countAmbiguity)
+ if (includeAmbiguity)
{
boolean ambiguity = isNucleotideAmbiguity(C);
if (ambiguity)
return false;
}
+ /**
+ * Tests *only* nucleotide ambiguity codes (and not definite nucleotide codes)
+ */
public static boolean isNucleotideAmbiguity(char c)
{
switch (Character.toUpperCase(c))
*/
public static boolean isNucleotideSequence(String s, boolean allowGaps)
{
+ return isNucleotideSequence(s, allowGaps, false);
+ }
+
+ public static boolean isNucleotideSequence(String s, boolean allowGaps,
+ boolean includeAmbiguous)
+ {
if (s == null)
{
return false;
for (int i = 0; i < s.length(); i++)
{
char c = s.charAt(i);
- if (!isNucleotide(c))
+ if (!isNucleotide(c, includeAmbiguous))
{
if (!allowGaps || !isGap(c))
{