RNA mode: How should template search be performed for rnaSequence inputs?

Hi,

I am currently running Protenix with an RNA-only input and have a question about template search in RNA mode.

My input JSON looks like this:

```json
{
"sequences": [
{
"rnaSequence": {
"sequence": "ACCAGGAUGGCCGAGUGGUUAAGGCGUUGGACUUAAGAUCCAAUGGACAUAUGUCCGCGUGGGUUCGAACCCCACUCCUGGUACCA",
"count": 1,
"unpairedMsaPath": "./examples/examples_with_rna_msa/9gmw/9gmw_2_all.a3m"
}
}
],
"name": "9gmw_2"
}
```

I noticed that the current template search implementation (`update_template_info` in `runner/template_search.py`) only checks for:

```python
if "proteinChain" in sequence:
```

So when the input contains only `rnaSequence`, the function does nothing and returns `False`.

I also noticed that the template search uses:

```
alphabet="amino"
```

which seems to indicate that the current template search pipeline is designed for protein sequences.

My questions are:

1. What is the intended way to perform template search for RNA-only inputs (`rnaSequence`)?
2. Should RNA template search be supported through the same `update_template_info` pipeline?
3. If RNA template search is supported, what database and pipeline should be used (e.g., RNA PDB templates)?
4. Or is the recommended approach to run RNA inference **without templates**?

Currently I tried providing:

```
seqres_database_path = pdb_seqres_NA.fasta
```

but since the code path is restricted to `proteinChain`, the template search is never triggered.

Could you clarify the recommended workflow for RNA template usage in Protenix?

Thanks!

Start in runner/template_search.py at update_template_info and trace the proteinChain guard and alphabet="amino" setting. Compare those paths with the rnaSequence input shown in the issue, then document whether RNA templates are supported, which database and pipeline apply, and whether template-free inference is the intended workflow.

Written by the indexing model from the issue text.