Inconsistent Chain IDs in results

Hi all,

I am having issues with inconsistent chain IDs in the prediction files. When I am folding homotrimers, the chain IDs are A, B, C in the first result, but then B, C, D, or A, B, B etc in the remaining ones.

Any suggestions as who to fix? I am using a custom made docker container built with micromamba instead of conda, git clone to build image is from June 28th 2023.

Thanks,

Lars

```
python run_alphafold.py \
--data_dir=$DBDIR \
--output_dir=$OUTDIR \
--uniref90_database_path=$UNIREFPATH90 \
--uniref30_database_path=$UNIREFPATH30 \
--uniprot_database_path=$UNIPROTPATH \
--mgnify_database_path=$MGNIFYPATH \
--max_template_date=2023-12-31 --template_mmcif_dir=$MCIFPATH \
--obsolete_pdbs_path=$OBSOLETEPATH \
--bfd_database_path=$BFDPATH --pdb_seqres_database_path=$PDBSQRES \
--use_gpu_relax=True \
--model_preset=multimer\
--fasta_paths=/data/input/some.fasta
```