Skip to content

PSB 2026

Pacific Symposium on Biocomputing 2026 — Fairmont Orchid, Kohala Coast, Big Island of Hawaii · January 3–7, 2026 (workshops Jan 3; scientific program Jan 4–7).

Status: ✅ built from the open proceedings. Unlike most of the corpus, this vault is not a scaffold and not transcript-derived — PSB publishes full peer-reviewed papers, free, before the meeting. All 55 papers, 5 session introductions, and 5 workshop introductions are open PDFs with PubMed records. Everything here is sourced from psb.stanford.edu; per-paper detail lives on the session pages.

  • Volume: Pacific Symposium on Biocomputing Vol. 31 — cited as Pac Symp Biocomput. 2026;31:<pages>
  • Edition: the 31st PSB, marking 30 years (PSB began in 1996)
  • Co-chairs: Russ B. Altman (Stanford), Lawrence Hunter (University of Chicago), Marylyn D. Ritchie (Medical University of South Carolina), Teri E. Klein (Stanford)
  • Coordinator: Tiffany Murray (Stanford Bioengineering), who has run peer review and proceedings assembly since 2001
  • Funding: NLM/NIH R13LM006766; NSF #2531429; ISCB travel support
  • Proceedings: psb.stanford.edu/psb-online/proceedings/psb26/ — free, CC BY-NC 4.0
  • Publisher: Biocomputing 2026, World Scientific — book DOI 10.1142/14631, print ISBN 978-981-98-2474-8

Two date discrepancies on the PSB site itself

The post-event index page says PSB 2026 "was held on January 4-8, 2026," but the official schedule PDF header reads January 3–7, 2026 on every page, the workshop page agrees, and the day/date pairs (Sunday Jan 4, Wednesday Jan 7) are consistent with the 2026 calendar. We use Jan 3–7. Similarly, the post-event page says "four workshops" — the preface, the proceedings TOC, and the schedule all show five, and five workshop introductions are published (pp. 849–868).

Why this is in the vault

PSB is the highest-yield harvest in the corpus, per unit of effort. Every other concluded meeting we track gives us abstracts, slides, or auto-generated captions. PSB gives complete, peer-reviewed, openly licensed papers — three referees each — indexed in PubMed within weeks. There is no embargo to wait out and no paywall to route around.

Three further reasons it earns a vault:

  1. Foundation models and LLMs are the dominant theme. At least 20 of 55 papers are LLM- or FM-centric, and unusually many of them are evaluations rather than model announcements — including a multi-institution negative result (Automated Chest X-ray Report Generation Remains Unsolved). That is exactly the critical-evaluation posture this corpus takes in Why linear baselines win and the evaluation papers catalog.
  2. It is the front edge of LLM agents in biology. Four papers apply agentic LLMs to biological tasks — gene-set analysis, protein function prediction, disease-relationship discovery, and clinical agent benchmarking — which is the corpus's agentic-ai thread with published methods attached.
  3. PSB's session model makes it a leading indicator. Sessions are proposed and selected fresh each year rather than fixed, so the track list is itself a signal about which subfields the community thinks are emerging. See Themes.

What PSB is not: there is no cancer session. Oncology enters through methods papers — see the cancer thread for the four that matter.

Program shape

PSB's format is unusual and worth knowing before reading the session pages. Each peer-reviewed session runs: a 10-minute organizer introduction (published as a paper), ~15-minute talks per accepted paper, then a full hour of open discussion. Around that sit five non-peer-reviewed workshops, invited "TERI Talks," a poster session, two keynotes, and the Research Parasite / Research Symbiont Awards.

Track Papers Page
S1 Precision Medicine: integrating large-scale data and intermediate phenotypes 10 →
S2 Biological Molecular Function: methods and benchmarks for biological dark matter 7 →
S3 Systems Biology and Network Analysis: from multi-omics integration to mechanisms 6 →
S4 AI and Machine Learning in Clinical Medicine 27 →
S5 Fairness and Bias in Biomedical AI/ML 5 →
Total peer-reviewed 55

Plus: Keynotes and invited talks · Workshops · Themes

The AI/ML in Clinical Medicine session is half the volume of the entire meeting — a fact worth reading as a statement about where the field's submissions are going.

What we have to work with

Source Coverage Notes
Online proceedings TOC all 65 records psb-online/proceedings/psb26 — free, last updated Dec 19, 2025
Per-paper PDFs all 65 psb.stanford.edu/psb-online/proceedings/psb26/<name>.pdf
Full proceedings single file master PDF — ~120 MB
Poster / abstract book 117 records HTML · XLSX — 83 numbered posters + 32 oral-no-poster, with full author lists, affiliations, 300-word abstracts, preprint DOIs. Highest-yield unharvested target.
Schedule full program schedule.pdf — the only source for invited-talk slots
Preface organizers, reviewers preface.pdf, pp. xi–xv
Keynote abstracts + bios 2 keynotes keynote page
Video keynotes at minimum YouTube playlist · index — contents not yet enumerated
PubMed 65 / 65 "Pac Symp Biocomput"[jour] AND 2026[dp]
PMC 22 / 65 author manuscripts, not publisher versions
Slides none found two workshops posted agenda PDFs only

Known data caveats

  • License is CC BY-NC 4.0, not CC BY. The proceedings TOC page claims CC BY 4.0, but the copyright line inside each paper and the PubMed record both say non-commercial. Treat as CC BY-NC.
  • One paper is unlinked from the TOC. cardone.pdf (Cardone et al., pp. 629–643) exists and returns HTTP 200 but has no link on the proceedings index — a link scrape will silently miss it. Direct URL: https://psb.stanford.edu/psb-online/proceedings/psb26/cardone.pdf
  • One page range conflicts. rifat.pdf is listed as pp. 644–665 on the TOC but 644–663 in PubMed; the next paper starts at 664, so PubMed is likely right.
  • A name is rendered two ways. The Fairness session organizer appears as "Madalena Ng" on the TOC and preface, "Magdalena Ng" in the session-intro byline.
  • No acceptance rate is published. None appears in the preface, post-event site, or schedule. Third-party sites quote a generic figure for PSB; do not cite it as a 2026 fact.
  • World Scientific blocks automated fetching (HTTP 403). ISBN and DOI facts here are cross-checked from the master-PDF filename, the chapter DOI prefix, and retailer listings rather than read off the publisher's copyright page.

What to harvest next

  1. The poster/abstract book. 117 records with affiliations and preprint DOIs, in both HTML and XLSX — structured, free, and not yet in the corpus. This is where the single-cell and spatial material lives (see Themes).
  2. The YouTube playlist. Confirm which talks were actually recorded; the site only commits to keynotes.
  3. Per-paper pages for the FM-evaluation cluster, matching the depth of the SCGD26 talk pages.

Sources