BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260611T145139Z
LOCATION:Bldg. 6 - Room 003
DTSTART;TZID=Europe/Stockholm:20260701T120000
DTEND;TZID=Europe/Stockholm:20260701T123000
UID:submissions.pasc-conference.org_PASC26_sess173_pap131@linklings.com
SUMMARY:Scalable Agentic Reasoning for Designing Biologics Targeting Intri
 nsically Disordered Proteins
DESCRIPTION:Matthew Sinclair, Moeen Meigooni, and Archit Vasan (Argonne Na
 tional Laboratory); Ozan Gokdemir (University of Chicago); Xinran Lian, He
 ng Ma, and Yadu Babuji (Argonne National Laboratory); Alexander Brace (Uni
 versity of Chicago, Argonne National Laboratory); Khalid Hossain (Argonne 
 National Laboratory); Carlo Siebenschuh (University of Chicago); Thomas Br
 ettin (Argonne National Laboratory); Kyle Chard (University of Chicago); C
 hristopher Henry, Daniel Schabacker, and Venkatram Vishwanath (Argonne Nat
 ional Laboratory); Rick Stevens and Ian Foster (Argonne National Laborator
 y, University of Chicago); and Arvind Ramanathan (Argonne National Laborat
 ory)\n\nIntrinsically disordered proteins (IDPs) represent crucial therape
 utic targets due to their significant role in disease–approximately 80% of
  cancer-related proteins contain long disordered regions – but their lack 
 of stable secondary/tertiary structures makes them“undruggable.” While rec
 ent computational advances, such as diffusion models, can design high-affi
 nity IDP binders, autonomous systems can accelerate their translation to p
 ractical drug discovery by reducing the need for expert intervention acros
 s large-scale design campaigns. To address this challenge, we designed and
  implemented StructBioReasoner, a scalable multi-agent system for designin
 g biologics that can be used to target both IDPs and structured proteins. 
 StructBioReasoner employs a novel tournament-based reasoning framework whe
 re specialized agents compete to\ngenerate and refine therapeutic hypothes
 es, naturally distributing computational load for efficient exploration of
  the vast design space. Agents integrate domain knowledge with access to l
 iterature synthesis, AI-structure prediction, molecular simulations, and s
 tability analysis, coordinating their execution on HPC infrastructure via 
 an extensible federated agentic middleware, Academy. We\nbenchmark StructB
 ioReasoner across Der f 21 and NMNAT-2 and demonstrate that over 50% of 78
 7 designed and validated candidates for Der f 21 outperformed the human-de
 signed reference binders from literature, in terms of improved in silico b
 inding free energy. For the more challenging NMNAT-2 protein, we identifie
 d three binding modes from 97,066 binders, including the well-studied NMNA
 T2:p53 interface. Thus, StructBioReasoner lays the groundwork\nfor agentic
  reasoning systems for IDP therapeutic discovery on Exascale platforms.\n\
 nDomain: Engineering, Life Sciences, Physics\n\nSession Chair: Andrei Onut
  (University of Basel)\n\n
END:VEVENT
END:VCALENDAR