Calculation of locus boundaries in rnazCluster.pl

[wash]

The extract of the following e-mail points out what could be a bug (at least I don't have an explanation for it at the moment)

The second question relates to the definition of the locus coordinates. In short, how are these extracted? In many (most) cases, the locus that comes out in the gff is the limit of the high scoring windows within the locus ... for example,

locus4247       scaff1   5683094 5683288 2       62.71   0.999651        -2.58
window7171      locus4247       scaff1   5683174 5683288 -       2       118     50.85   -33.90  -21.05  -19.05  -2.00   1.55    -2.58   0.62    3.84    0.999651
window7170      locus4247       scaff1   5683134 5683250 -       2       118     58.47   -33.15  -22.73  -21.73  -1.00   1.44    -1.89   0.69    2.43    0.993843
window7169      locus4247       scaff1   5683094 5683212 -       2       118     62.71   -40.51  -24.83  -24.33  -0.50   1.33    -2.21   0.61    1.82    0.978602

generates:

scaff1 RNAz structuredRNA 5683095 5683288 0.999651 . id "locus4247"


however, in other cases I see things like:

locus1690       scaff1   2543458 2543651 2       62.83   0.999949        -2.41
window2799      locus1690       scaff1   2543458 2543571 -       2       113     58.41   -40.20  -24.40  -22.90  -1.50   1.47    -2.16   0.61    1.98    0.984572
window2801      locus1690       scaff1   2543498 2543611 +       2       113     62.83   -41.75  -34.00  -34.25  0.25    1.30    -2.41   0.81    4.78    0.999949
window2800      locus1690       scaff1   2543498 2543611 -       2       113     62.83   -39.05  -26.15  -30.65  4.50    1.12    -1.79   0.67    1.45    0.954832

generating the gff entry:

scaff1 RNAz structuredRNA 2543459 2543651 0.999949 . id "locus1690"


I am a bit confused as to where the coordinate "2543651" is coming from. 

[svenderheld]

If someone runs into that issue please post an minimal input file to reproduce the described behaviour.