Biology subjects

Brimo, n.

Publications and source records attributed to Brimo, n..

2 recordsLinked to original sources

bioRxiv2026

Hierarchical Temporal Transformer for Cancer Grade Prediction and Cross-Cancer Transfer Learning from Pathology Reports

Background: Language models for cancer reports usually read one document at a time and are tested on tumour types seen in training. Continuous-time sequence models address the first gap, but whether they use visit timing, not just order, is seldom tested. Methods: We developed the Hierarchical Temporal Transformer (HTT), which encodes reports with adapted BiomedBERT and reads the sequence with a cancer-type-conditioned temporal transformer. We evaluated it against best-possible AUROC ceilings on two synthetic benchmarks, one with the label stated in the text and one recoverable only from tumour sizes and visit intervals (ten seeds). Diagnostics led to an interval-aware time code (HTT-IA). On 4,786 TCGA pathology reports in 14 types, three types were withheld, and a text-only HTT was compared with a survival model. Results: On the phrase benchmark HTT reached AUROC 0.931 +/- 0.012 against 0.913 +/- 0.007 for a last-report baseline (+0.018, 95% CI -0.009 to +0.050); dates cannot raise the ceiling there. On the growth benchmark the initial time code scored 0.790, equal to a no-time model (0.787), whereas HTT-IA reached 0.843 (+0.053, 0.026 to 0.082) against ceilings of 0.894 without and 0.979 with dates. On unseen tumour types HTT reached 0.925 +/- 0.007, equal to in-distribution performance and to a single-report baseline. Text-only macro C-index was 0.592 (CALM 0.626-0.652). Conclusion: Cross-type transfer is carried by the pre-trained encoder. Visit timing is used only when the data make it identifiable and the time code is on a clinical scale. Ceilings, order-only ablations and re-dating probes are inexpensive checks for such models.

bioinformatics↗

bioRxiv2026

Optimiser Repairs for Living-Cell Self-Driving Laboratories: Performance and Redundancy under Experimental Constraints

Living-cell self-driving laboratories must optimise experiments despite delayed readouts, changing responses and limited sample availability. In simulations, we evaluate when optimiser repairs improve closed-loop performance and when platform constraints make them redundant, separating consumption of a sample from permanent loss of the ability to repeat its condition. We prove that exclusions and confirmation requests produce exactly the baseline trajectory when they act only on conditions already forbidden by the platform, share the same posterior and selection procedure, and release infeasible requested slots. This result covers the evaluated pending-exclusion, previous-batch and confirmation rules under permanent condition retirement; it does not cover arbitrary selection interventions. We then conduct 33,900 simulated policy runs with 200-300 paired campaign seeds per setting. A public PBMC reference supports characterisation of the target-label readout, while condition-response landscapes remain synthetic. Across 27 landscape, assay-size and kernel settings, previous-batch exclusion changes final normalised performance by -0.004 to +0.283, with positive differences surviving Holm adjustment in 13 settings. Effects are small on the rugged landscape. Permanent retirement can be harmful under noisy rugged responses. All three subsumed interventions reproduce every saved baseline event on 200 of 200 paired seeds. These results establish the importance of explicit availability constraints and evaluation across experimental regimes when assessing optimiser repairs.

bioinformatics↗