Custom-column extraction across hundreds of nanomedicine papers

Amal is a postdoctoral research fellow at the University of Queensland working in cancer nanomedicine and neuroscience. She formulates nanoparticles and tests them in vitro and in vivo, and supervises master’s students on nanomedicine projects each semester.

Her research spans materials science, pharmacology, and biomedical engineering, and the literature is scattered across all three. Reading hundreds of papers on biocompatibility and safety for regulatory compliance, and organizing that evidence for submissions, competed with her lab work and her students.

Paperguide became the filter between “hundreds of papers” and “the ones that matter for this question,” letting her compare studies on the specific dimensions she cares about rather than reading each in full.

Amal’s problem is one of scale meeting specificity. Her work on targeted drug-delivery nanoparticles for cancer means that for any given question, the answer is spread across hundreds of papers, and what she needs from each one is narrow and technical: not the gist, but the toxicity profile, the clearance rate, the exact delivery mechanism. Reading hundreds of papers in full to pull four data points from each was never sustainable alongside running a lab and supervising students.

Instead, she turns the questions she cares about into columns. She sets up an extraction with fields like toxicity profile, clearance rate, biocompatibility, and delivery mechanism, points it at her collection of hundreds of nanoparticle studies, and gets back a table where each study is a row and each of her questions is answered, cell by cell, from the paper’s own text. From there, selecting the relevant papers is a matter of scanning a table rather than reading a stack. When she needs to go deeper on a specific study’s safety data, she asks it directly, “What were the toxicity profiles and clearance rates reported?”, and pulls the specifics she needs for experimental design or a regulatory file. Because every value in that table traces back to its source, the safety and biocompatibility evidence she assembles for regulatory submissions is organized and defensible from the start, not a spreadsheet she has to justify later.

The effect is dramatic on the task that used to be pure grind: reviewing specific information across a hundred papers dropped from around thirty minutes to about five. That reclaimed time goes back where she wants it, to experimental design and to her students.