Rosetta RNA_Denovo reports “Not complementary” for valid A–U pair – format issue suspected

[fxwt77777]

Environment:

Rosetta version: source release 3.14 (2024.03)

Executable: rna_denovo.default.linuxgccrelease

OS: Ubuntu 24.04.2 LTS (running via WSL2)

Python: 3.13.2

RNAfold version: 2.4.7

Command:

rna_denovo.default.linuxgccrelease
-fasta aptamer_0.fasta
-secstruct_file aptamer_0.dbn
-nstruct 1
-out:file:silent aptamer_0.out
-out:file:silent_struct_type binary
-minimize_rna true
-database [path to rosetta database]

Error encountered:

ERROR: Not complementary at positions U 4 and A 21
ERROR:: Exit from: src/core/pose/rna/RNA_SecStruct.cc line: 284

What I’ve checked:

Sequence in .fasta and structure in .dbn are both 60 nt long.

The base at position 4 is U, at position 21 is A → canonically complementary.

All base pairs were verified using a script and match standard Watson-Crick or GU wobble pairs.

All brackets in .dbn are matched.

No leading/trailing spaces or encoding issues visible in files.

Similar errors arise for other sequences at different base-pair positions, even when they appear valid.

Example snippet:

aptamer_0
UGCUCCCGCGGUGCGGGUUAAGCAACCCAUCCUCAGGUACCAAAUGGGUGAGGAGGAAAA

dbn
(((((((((...)))))...))))..((.(((((....(((.....))))))))))....

Question:

Could Rosetta be failing on formatting quirks (e.g., newline, encoding, buffer issue)?

Are there strict constraints (e.g., min loop length or pairing distance) that I’m missing?

How can I debug what Rosetta sees internally at positions 4 and 21?

Any help would be appreciated — this is part of a pipeline to generate structures for machine learning training.

Thanks!

[roccomoretti]

This looks to be an uppercase/lowercase issue. The data structure it's using to check for complementarity needs lower-case letters. a is complementary to u (and t), but A is complementary to nothing.