Pseudoknot 90 lab ‎“fatal error : pair length does not match sequence length”

(Late) following up on this: It appears to be due to NuPACK running out of memory. NuPACK’s pseudoknot folding (in particular) is highly inefficient, and gets progressively worse as the length increases. At ~180 bases, for example, NuPACK can use 3-4GB (!) of RAM (in addition to taking a pretty long time to fold). This can particularly be an issue on mobile devices for example (in many cases the PK90 would crash immediately), if have many browser tabs or other programs open (particularly on lower-memory devices like Chromebooks), etc. Unfortunately this isn’t something we have control over (or can even reliably detect), so I’ll be marking this issue as “wontfix”. We’ve recently introduced a new engine that is significantly more efficient (and from what we’ve seen, more accurate!) which will be available going forward.