{
  "id": 47554,
  "title": "Optimized Multiple Circular Sequence Alignment for Cyclic Peptide Motif Discovery",
  "url": "https://urgent.news/2026/08/01/optimized-multiple-circular-sequence-alignment-for-cyclic-peptide",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-01T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.07.28.741376v1?rss=1"
  },
  "original_language": "en",
  "account": "The discovery of head-to-tail (H2T) cyclized peptides in drug development is gaining traction, as these peptides exhibit high target affinity, selectivity, and metabolic stability. Despite their chemistry being akin to linear amino acid chains, their linear sequence representation remains the standard input format for mainstream sequence generative models, which are instrumental in de novo peptide design (Slough et al., 2018; Rettie et al., 2025a;b). To identify conserved motifs that drive a family's function across a collection of such peptides, a multiple sequence alignment (MSA) is necessary. However, the complexity increases due to the cyclic nature of these peptides, which can be linearized at any residue, introducing the multiple circular sequence alignment (MCSA) problem. Current MCSA heuristics, while effective for genomic applications with a few tens of long sequences, prove to be inefficient for the cyclic peptide library scenario involving hundreds to thousands of shorter sequences. To address this gap, researchers have identified quality-preserving optimization opportunities, such as a library-scale preset tailored to short-sequence inputs, and added an orthogonal multi-core and SIMD backend for enhanced performance tuning. This approach results in near-linear thread scaling on the pairwise-comparison stage. The pipeline was validated using a library of 1,000 H2T cyclized peptides, each 18 residues long, targeting the oncoprotein Mouse double minute 2 human homolog (MDM2) generated by an internal peptide-design engine. Remarkably, the optimized MCSA maintained the same accuracy as the original implementation while achieving a remarkable improvement in speed, running over 650 times faster. This optimized MCSA tool is now publicly accessible at https://github.com/IVB-Generative-Biology/mars-turbo, enabling library-scale cyclic peptide sequence alignment.",
  "summary": "Head-to-tail (H2T) cyclized peptides are an increasingly important modality in drug discovery, combining high target affinity and selectivity with metabolic sta-bility. Because their underlying chemistry is still that of a linear amino-acid chain, their linear sequence representation is the native input format of main- stream sequence generative models now driving de novo peptide design (Slough…",
  "key_points": [
    "Optimized MCSA improves speed over 650 times for 1,000 H2T cyclized peptides",
    "Optimized approach maintains same accuracy as original implementation",
    "Tool publicly accessible at https://github.com/IVB-Generative-Biology/mars-turbo"
  ],
  "editors_take": null,
  "illustration": "https://urgent.news/ill/47554.png",
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}