geneid source docs

Description:

Read the external file containing HSPs from sequence alignments. Blast High-scoring Segment Pairs over the input sequence are projected over the input sequence, taking the best value in every position. Positions without homology support are assigned the value NO_SCORE (parameter file). By using these homologous regions to the input DNA sequence, exons overlapping any HSP will be enhancered and better predictions will be obtained.

HSPs file is in gff format (8 mandatory records plus string group optional).
If blastn (DNA vs DNA) is used to obtain the homologous regions, their frame is unknown, so a dot (".") can be used to make three copies from every one.

-- It is NOT necessary to provide the HSPs records ordered by any position --

Briefing:

packHSP* SelectHSP(packExternalInformation* external,
                   char* Locus,
                   long LengthSequence)

Select the group of HSPs to be integrated into the predictions according to current Locus name (using a hash table - dictionary of locus). Both FWD and RVS HSPs will be sorted using a quicksort routine.

long ReadHSP (char* FileName, 
              packExternalInformation* external)

To place every HSP into the correct array according to their blast frame (1,2,3) and strand. If frame is unknown (blastn), then 3 copies of the same HSP will be produced. (It IS necessary a previous sorting in the HSP file)