Random Protein Sequences in Seqqio: Seeds, Residue Weights, and FASTA Batches
Create a repeatable protein FASTA batch, control relative amino-acid weights, and verify what a seed reproduces. Follow executed examples and choose between new sampling and composition-preserving shuffling.
A random FASTA file is easy to create and harder to reproduce when its settings are missing. This tutorial creates 100 proteins of 200 amino acids, repeats the run, and then changes the sampling weights. You can inspect the resulting sequences and keep their generation settings in Seqqio's local history rather than reconstructing the experiment from a filename.
Build a protein FASTA batch you can recreate
Open Random Protein in Seqqio and select the batch generation mode. Set Fixed length to 200 amino acids and Sequence count to 100. Keep all 20 residues enabled under Uniform, enter seqqio_test as the title prefix, set the seed to 42, and enable unique results. Select Generate protein batch, then save the complete FASTA output.
| Setting | Value | Purpose |
|---|---|---|
| Length / count | 200 amino acids / 100 sequences | 20,000 residues across the batch |
| Distribution | Uniform; all 20 residues enabled | Equal sampling weight for each residue |
| Seed | 42 | Explicit repeatable random input |
| Unique results | Enabled | Reject duplicate complete sequence strings |
| Title prefix | seqqio_test | Stable, ordered FASTA identifiers |
The executed Seqqio engine returned 100 distinct sequence strings, each exactly 200 residues long, with zero duplicates discarded in this run. Every symbol belonged to the standard amino-acid alphabet. The first FASTA record was:
>seqqio_test_0001
FLPRNHPFYGEYFYYRVAKDIRVASMRMDDSPIQWKYPTEWPAVTNDTEAMRCSYIELHW
QKPSVYNFGKCKMVPFVKKTRGFGKFQMPNMEQHFHTTDIWMGKAGMCKEEDEKCFVGFV
ETEEEWRGKCSWEKFCGNTYEKNDFREWVGVMWTRPMTHGSGFADDGPNYQTYIARHMMP
THMQHGEIDDCKYFYKQHIRThese are software test sequences, not proteins drawn from an organism or curated database. The batch can help check whether a parser retains all 100 records, whether a downstream tool handles 200-residue inputs, and whether an exported result can be traced back to its original identifier.
Repeat the seed before changing the model
Run the same settings again with seed 42. In our executed control, the complete FASTA file was byte-for-byte identical. Changing only the seed to 43 produced a different file. That observation demonstrates this replay; it does not mean every possible seed change must produce a different sequence.
Seqqio reports its generation profile as ChaCha20-Seqqio-v1. Save the application version, profile, resolved seed, enabled residues or weights, length settings, count, uniqueness choice, and title prefix with the exported file. A seed alone is incomplete provenance. The Rust Rand reproducibility documentation explains why deterministic outputs can depend on the algorithm and library version. Do not assume an unrelated generator, or a future changed profile, will reproduce this batch.
If you leave the seed empty, Seqqio resolves one from system entropy and retains that value in history. Record the resolved seed when you want a repeatable control. Use New seed when you want to clear the current seed and request a fresh one on the next run.
Use relative weights to change amino-acid sampling
Switch to Custom weights. Keep the batch length and count unchanged, set A to 5, G to 3, C to 0, and each of the other 17 residues to 1. Keep seed 42 and unique results enabled. The total weight is 25: A has a sampling probability of 5/25, G has 3/25, and C is excluded.
| Residue | Relative weight | Sampling probability | Observed in 20,000 residues |
|---|---|---|---|
| A | 5 | 20% | 4,004; 20.02% |
| G | 3 | 12% | 2,374; 11.87% |
| C | 0 | 0% | 0; 0% |
| Each other residue | 1 | 4% each | Counts vary with the sampled batch |
The observed A and G percentages are close to their sampling probabilities, but the generator does not enforce exact residue counts. A single 200-residue sequence can differ more noticeably from those probabilities than the pooled 20,000-residue batch. Zero weight is different: C cannot be sampled under these settings. We independently counted the exported residues to confirm that exclusion.
Choose new sampling or composition-preserving shuffling
Use Random Protein when you want new sequences drawn from declared residue weights. If your control must retain the exact counts in an existing protein, use the Seqqio Shuffle Sequence tutorial. Shuffling rearranges the input symbols; weighted generation draws new symbols. Neither choice automatically supplies an appropriate biological null model. Decide which properties your analysis requires the control to preserve.
The general random-protein guide introduces synthetic sequence testing. This Seqqio workflow adds an exposed seed, explicit distribution settings, batch records, and saved local runs for repeatable work.
Export the complete batch and understand uniqueness
Use the FASTA or TSV save controls for the complete generated dataset. A truncated screen preview is not a truncated saved batch; when the interface offers Copy preview, copying is limited to the displayed subset. For a downstream test file, save the full export and check its record count rather than assuming everything visible is everything generated.
Unique results means distinct complete sequence strings, not biologically diverse or unrelated proteins. With only A enabled and a fixed length of eight, there is only one possible sequence: AAAAAAAA. Our control requested three unique sequences and received unique_sequence_space_too_small. Allowing duplicates returned three copies. Increasing the count cannot create diversity that the allowed sequence space does not contain.
When Seqqio fits this workflow
Random Protein is useful for repeatable fixtures, classroom exercises, and exploratory controls with a stated independent-residue sampling model. It does not reproduce natural sequence families, evolutionary relationships, motifs, structure, or expression behavior unless your separate study establishes that connection. The local workflow is a practical fit when you want to generate, inspect, save, and revisit batches without maintaining a generation script.
Random Protein is included in Seqqio's 39-application Windows 64-bit toolkit. The current offer is a US$99 one-time purchase with no activation key. Review the platform requirements and final checkout total on the product page.
References
- Rust Rand project contributors. Reproducibility The Rust Rand Book Official context for algorithm and version dependence; not a cross-version Seqqio guarantee.