search

veg / hyphy

HyPhy: Hypothesis testing using Phylogenies
http://www.hyphy.org
Other
200 stars 68 forks source link

Slight difference in GARD breakpoint location: screen output vs screened data #1679

Closed katherinermartin closed 3 months ago

katherinermartin commented 5 months ago

Hi! Apologies if this has been answered previously.

When GARD identifies breakpoints in my alignment, I consistently see a slight discrepancy in the breakpoint identified on the GARD results page (e.g., Figure 2. Model-averaged support for breakpoint placement) and the breakpoint identified when I download the partitioned results in the screened_data.nex file.

Attached is an example (both .json and the screened .txt); on-screen, the breakpoint is 183/184 bp, but in screened_data.nex, it's 184/185:

CHARSET span_1 = 1-184 CHARSET span_2 = 185-597

If I decide to manually partition, which breakpoint should I use? I feel like I'm missing something fairly obvious since they're so close to one another, so thank you in advance for any help :)

result_3rate_beta_gamma.json screened_data_3rate_beta_gamma.txt

spond commented 5 months ago

Dear @katherinermartin,

It may actually be a bug (adding an extra 1 to the NEXUS CHARSET); I'll take a closer look. In any case, for selection analyses HyPhy will move breakpoints to the nearest codon boundary.

Best, Sergei

github-actions[bot] commented 3 months ago

Stale issue message