Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AptaFind: A lightweight local interface for automated aptamer curation from scientific literature

Geoffrey Taghon

TL;DR

Aptamer literature is scattered and manual searches are time-consuming. AptaFind couples on-device semantic understanding with deterministic parsing in a three-tier output—direct sequence extraction, curated leads, and exhaustive literature discovery—without cloud dependencies. Validation on the UT Aptamer Database shows Tier 3 and Tier 2 reaching $84.0\%$ and Tier 1 reaching $79.3\%$, with a practical throughput of about $954\pm 43$ targets/hour on a Mac Studio, evidencing efficient, scalable performance. The Minimum Agentic Flow principle—balancing semantic processing with reliable data handling—enables actionable scientific intelligence across domains while preserving privacy, and the open-source release promotes broader adoption.

Abstract

Aptamer researchers face a literature landscape scattered across publications, supplements, and databases, with each search consuming hours that could be spent at the bench. AptaFind transforms this navigation problem through a three-tier intelligence architecture that recognizes research mining is a spectrum, not a binary success or failure. The system delivers direct sequence extraction when possible, curated research leads when extraction fails, and exhaustive literature discovery for additional confidence. By combining local language models for semantic understanding with deterministic algorithms for reliability, AptaFind operates without cloud dependencies or subscription barriers. Validation across 300 University of Texas Aptamer Database targets demonstrates 84 % with some literature found, 84 % with curated research leads, and 79 % with a direct sequence extraction, at a laptop-compute rate of over 900 targets an hour. The platform proves that even when direct sequence extraction fails, automation can still deliver the actionable intelligence researchers need by rapidly narrowing the search to high quality references.

AptaFind: A lightweight local interface for automated aptamer curation from scientific literature

TL;DR

Aptamer literature is scattered and manual searches are time-consuming. AptaFind couples on-device semantic understanding with deterministic parsing in a three-tier output—direct sequence extraction, curated leads, and exhaustive literature discovery—without cloud dependencies. Validation on the UT Aptamer Database shows Tier 3 and Tier 2 reaching and Tier 1 reaching , with a practical throughput of about targets/hour on a Mac Studio, evidencing efficient, scalable performance. The Minimum Agentic Flow principle—balancing semantic processing with reliable data handling—enables actionable scientific intelligence across domains while preserving privacy, and the open-source release promotes broader adoption.

Abstract

Aptamer researchers face a literature landscape scattered across publications, supplements, and databases, with each search consuming hours that could be spent at the bench. AptaFind transforms this navigation problem through a three-tier intelligence architecture that recognizes research mining is a spectrum, not a binary success or failure. The system delivers direct sequence extraction when possible, curated research leads when extraction fails, and exhaustive literature discovery for additional confidence. By combining local language models for semantic understanding with deterministic algorithms for reliability, AptaFind operates without cloud dependencies or subscription barriers. Validation across 300 University of Texas Aptamer Database targets demonstrates 84 % with some literature found, 84 % with curated research leads, and 79 % with a direct sequence extraction, at a laptop-compute rate of over 900 targets an hour. The platform proves that even when direct sequence extraction fails, automation can still deliver the actionable intelligence researchers need by rapidly narrowing the search to high quality references.
Paper Structure (19 sections, 1 figure)

This paper contains 19 sections, 1 figure.

Table of Contents

  1. Introduction
  2. System Design
  3. Multi-Source Literature Discovery
  4. Enhanced XML Processing and Sequence Extraction
  5. Deterministic Validation and Deduplication
  6. Three-Tier Output Generation
  7. Results and Discussion
  8. Benchmarking
  9. Limitations and Future Directions
  10. Broader Impact
  11. Methods
  12. Database Construction from PDFs
  13. Online Literature Search Implementation
  14. Enhanced XML Processing and Validation
  15. Sequence Processing Algorithms
  16. ...and 4 more sections

Figures (1)

  • Figure 1: AptaFind implements a three-tier research intelligence approach ensuring every search delivers actionable value.(A) System Architecture: Multi-source literature discovery combines PubMed, PMC, and bioRxiv searches with supplement harvesting and browser automation for comprehensive coverage. Enhanced XML processing extracts structured metadata while deterministic sequence validation ensures reliability. The three tiers deliver calibrated intelligence: Tier 1 (direct sequence extraction, "diamond quality"), Tier 2 (curated research leads with complete metadata, "diamonds in the rough"), and Tier 3 (searched exhaustively but found no leads with metadata). Testing three random subsets of 100 targets from the UT Database on a consumer desktop demonstrates (B) processing speed of $\sim$953 targets per hour ($\sim$4 seconds/target) and (C) hit rates across all tiers: $\sim$84 % literature discovery, $\sim$84 % research lead generation, 79 % direct sequences returned.