Hi guys,

I've been analysing a data set of 5 protein-coding genes (most of them are not the complete sequences), and found out this strange behaviour of HyPhy when trying to read some of them (via a batch file): sequences starting with reading frame "231" - which coincidentally also happen to be ([divisible by 3 ]+ 1) bases long in my data - are missing 1bp when read into the 3 different codon-position partitions, so that each category has exactly the same number of bases. Due to that, the analyses are returning slightly different ML values, partition weights and branch length estimates when compared to PAUP.

How can I fix that? Should I just discard one base and stick to that, regarding those genes?

Regarding the other reading frames:

"312", ([div. by 3]+ 2 bases long), 1 gene - everything is normal.
"123", ([div. by 3] bases long), 2 genes - everything is normal.


Jose_Patane

As always, thks again, your code worked out - branches and ML estimates are ok now!

Anyway, for future analysis I'd like to know what I've done wrong, so here's how I was doing it:

- position of "gene_231" (meaning the 1st codon position starts on site #2) in the concatenated sequence of all the genes (not 0-indexed, this is actually the raw position in the alignment):  2741-3119 (379 bp).

Here's the code I had formerly used to partition the data into codon positions:

------------------------------------------------------------------
DataSet All_seqs= ReadDataFile("my_concatenated_data");

DataSetFilter gene_231 = CreateFilter (All_seqs,1,"2740-3118");

DataSetFilter filteredData1 = CreateFilter (gene_231,1,"<010>");
DataSetFilter filteredData2 = CreateFilter (gene_231,1,"<001>");
DataSetFilter filteredData3 = CreateFilter (gene_231,1,"<100>");

   .....
   .....
   .....

fprintf (stdout,"n_sites_1st = ", filteredData1.sites);
fprintf (stdout,"n_sites_2nd = ", filteredData2.sites);
fprintf (stdout,"n_sites_3rd = ", filteredData3.sites);

   .....
   .....
   .....

-------------------------------------------------------------------------

... with results:

"n_sites_1st = 126
n_sites_2nd = 126
n_sites_3rd = 126"


Like I had said, the starting sequences at '312' and '123' return proper results using this kind of code... can you help me figuring out what went wrong?

Once more, thank you Sergei, problem resolved!

Jose_Patane

P.S.: I changed the title of my first report in this topic, it was not a bug of the program after all... I'll be more careful next time!