GenoMorph: Pathway-Grounded Genomic Disease Reasoning via Adaptive Latent Computation

Tanmoy Kanti Halder; Akash Ghosh; Arijit Roy; Sriparna Saha

cs.AISep 28, 2026

GenoMorph: Pathway-Grounded Genomic Disease Reasoning via Adaptive Latent Computation

Authors: Tanmoy Kanti Halder, Akash Ghosh, Arijit Roy, Sriparna Saha

Organizations: Department of Computer Science and Engineering, Indian Institute of Technology Patna, Bihta, 801106, Bihar, India · Prasannadeb Women’s College, India

Abstract

Large language models (LLMs) have demonstrated strong capabilities in biological reasoning; however, genomic disease inference remains largely dependent on memorized gene-disease associations rather than understanding biological pathways. This shortcut learning undermines robustness and generalization, and breaks down when molecular identifiers are unavailable. We present GenoMorph, a multimodal genomic reasoning framework that shifts disease prediction from associative gene-disease mapping toward pathway-grounded reasoning. GenoMorph couples a frozen DNA foundation model with question-conditioned cross-attention fusion, self-adaptive latent reasoning (LatentSp), a residual reasoning gate for iterative genomic evidence reinjection, and rejection sampling fine-tuning regularized by hierarchical optimal transport (OT). Rather than learning direct gene-disease mappings, GenoMorph aligns genomic sequence representations with latent pathway dynamics, enabling reasoning trajectories that follow molecular interactions before producing disease predictions. LatentSp dynamically allocates computation according to reasoning confidence, reducing unnecessary reasoning steps and improving inference efficiency. We further construct an anonymized benchmark from the Kyoto Encyclopedia of Genes and Genomes (KEGG), replacing every gene and molecular identifier with anonymous symbols while preserving sequences and pathway topology, thereby removing memorization shortcuts. GenoMorph raises the weighted F1 from 0.7863 (BioReason) to 0.9412, and rejection sampling fine-tuning with self-adaptive latent reasoning pushes it to 0.9725 while cutting latency nearly 60%. On the anonymized benchmark it reaches 0.9465 F1, substantially outperforming prior systems and confirming that accurate disease prediction can arise from pathway reasoning rather than memorized gene-disease associations.

Figures & tables

Figure 1: Training and inference pipeline of GenoMorph . The framework begins by encoding genomic DNA sequences using the frozen Evo2-7B foundation model to obtain DNA embeddings, which are combined with clinical question embeddings through question-conditioned CrossAttentionFusion (Stage 1) to generate question-aware genomic representations. Then it performs curriculum-based LatentSp training to initialize latent reasoning, followed by ThinkingResidualGate pretraining for controlled genomic evidence reinjection. Stage 2 computes offline Hierarchical Optimal Transport (HiRef-OT) target manifolds and correspondences ( $d_{\mathrm{OT}}$ ), providing geometry-aware rewards for reinforcement optimization. Stage 3 first applies GRPO to optimize latent reasoning trajectories, genomic evidence reinjection, and language generation using the HiRef-based reward. The resulting reasoning traces are subsequently filtered and used for Rejection Sampling Fine-Tuning (RSFT), enabling the model to internalize the adaptive reasoning policy without requiring explicit per-step entropy estimation or a learned entropy threshold during inference. The final self-adaptive latent model dynamically allocates computation by continuing latent reasoning for difficult samples while reinjecting genomic evidence only when necessary, producing open-ended disease predictions grounded in biologically coherent pathway reasoning rather than direct gene–disease associations.

Component	Setting
Backbone LLM	Qwen3-1.7B-Instruct
Genomic encoder	Evo2-7B (frozen)
DNA embedding layer	blocks.28.mlp.l3
Fusion module	CrossAttentionFusion
Trainable parameters	CrossAttentionFusion + LoRA adapters
LoRA rank / $\alpha$ / dropout	32 / 64 / 0.05

Table 1: Training configuration for supervised learning (Stage 1).

Parameter	Value
Initialization	Best supervised checkpoint
Optimizer	AdamW
Learning rate	$2\times 10^{-6}$
Batch size	1 (effective = 4)
LoRA ( $r$ , $\alpha$ )	16, 32
Generations	8

Table 2: Training configuration for Group Relative Policy Optimization (GRPO) (Stage 3).

Parameter	Value
Initialization	Best GRPO checkpoint (LoRA merged)
Optimizer	AdamW
Learning rate	$1\times 10^{-5}$
Weight decay	0.01
Learning rate schedule	Cosine decay (5% warmup)
Batch size	1 (effective = 8)

Table 3: Rejection Sampling fine-tuning (RSFT) configuration.

Category	Model	Input	Setting
General	GPT-4o	Text	Zero-shot
General	Gemini	Text	Zero-shot
General	Qwen3-1.7B	Text	Zero-shot
Biomedical	BioMedGPT	Text	Zero-shot
Biomedical	BioMistral	Text	Zero-shot
Biomedical	Meditron	Text	Zero-shot

Table 4: Baseline methods used for comparison.

\multirow 2* Model	Named Benchmark				Anonymous Benchmark
	Acc.	Macro-F1	Weighted-F1	Time (s)	Acc.	Macro-F1	Weighted-F1	Time (s)
Trained Models
LLM-only (Qwen3)	0.8897	0.5959	0.7461	21.45	0.4931	0.1921	0.4256	25.93
BioReason (SFT)	0.9069	0.7700	0.7863	23.74	0.6034	0.3798	0.6032	28.83
BioReason (GRPO)	0.8275	0.6828	0.8550	25.91	0.5172	0.2176	0.4812	25.37
GenoMorph (Ours)	0.9552	0.8461	0.9412	18.31	0.9207	0.8266	0.9285	23.64

Table 5: Overall performance on the genomic disease reasoning benchmark. Results are reported on both the original benchmark (Named) and the anonymized benchmark (Anonymous). Macro-F1 (M), Weighted-F1 (W), and average inference time per example are reported. Best results are shown in bold .

Figure 2: Accuracy comparison on the named and anonymized genomic disease reasoning benchmarks. The vertical distance between the paired markers represents the performance degradation caused by removing explicit gene and molecular entity names. GenoMorph exhibits the smallest performance gap, demonstrating robust sequence-level genomic reasoning under entity anonymization.

Figure 3: Average inference time per sample on the named and anonymized genomic disease reasoning benchmarks. Lower inference time indicates higher computational efficiency. GenoMorph with self-adaptive RSFT achieves the lowest inference latency while simultaneously providing the highest predictive performance.

Model	$\Delta$ Accuracy $\downarrow$	Accuracy Retention (%) $\uparrow$
LLM-only (Qwen3)	39.66	55.4
BioReason (SFT)	30.35	66.5
BioReason (GRPO)	31.03	62.5
GenoMorph (Ours)	3.45	96.4
GenoMorph+ (Self-Adaptive RSFT)	2.76	97.2

Table 6: Robustness to gene-name anonymization. Absolute decrease in accuracy and accuracy retention after replacing all gene and molecular entity names with globally consistent anonymous identifiers. Lower accuracy drops and higher retention indicate stronger sequence-level genomic reasoning.

Figure 4: Anonymous benchmark accuracy and accuracy retention after gene-name anonymization. Gray bars denote baseline models, whereas blue bars correspond to the proposed GenoMorph variants. The red line represents accuracy retention relative to the named benchmark. GenoMorph and GenoMorph +RSFT achieve both substantially higher anonymous accuracy and significantly greater accuracy retention, demonstrating robust sequence-level genomic reasoning under entity anonymization.

\multirow 2* Variant	\multirow 2* DNA Fusion	\multirow 2* LatentSp	\multirow 2* Latent Reasoning	\multirow 2* Gate	\multirow 2* HiRef OT	Named Benchmark			Anonymous Benchmark
						Acc	Weighted-F1	Time (s)	Acc	Weighted-F1	Time (s)
Stage 1: CrossAttn SFT	CrossAttn	✗	✗	✗	✗	0.9310	0.7573	43.85	0.5586	0.4762	26.84
Stage 1.5: +LatentSp	CrossAttn	✓	✗	✗	✗	0.9552	0.9538	15.05	0.9034	0.8952	17.10
Stage 1.5: +Gate	CrossAttn	✓	✗	✓	✗	0.9828	0.9793	14.78	0.9069	0.9073	17.52
GRPO ( $\theta_{\mathrm{low}}=0$ )	CrossAttn	✓	No Latent	✓	✓	0.9379	0.8690	29.24	0.7518	0.7074	30.82
GRPO ( $\theta_{\mathrm{low}}=3$ )	CrossAttn	✓	Full Latent	✓	✓	0.9448	0.9067	17.11	0.7459	0.7362	18.45

Table 7: Ablation study of GenoMorph . The table shows the contribution of each architectural component, including Cross-Attention DNA fusion, LatentSp curriculum learning, latent reasoning, adaptive gating, HiRef OT reward, and self-adaptive RSFT.

Figure 5: Optimization trajectory of GenoMorph throughout the ablation study in the accuracy–efficiency plane. Each point corresponds to one training stage or architectural variant. The blue dashed curve denotes the optimization trajectory as successive components are incorporated, while the red solid line represents the Pareto frontier. The red dashed segment illustrates the performance recovery obtained by introducing the HiRef OT reward compared with its ablated counterpart. The shaded green region indicates the desirable operating regime with simultaneously high prediction accuracy and low inference time. The final self-adaptive RSFT model achieves the best trade-off between predictive performance and computational efficiency.

Figure 6: Qualitative comparison on the original benchmark. GenoMorph produces biologically grounded reasoning and correctly predicts the target disease, whereas GPT-4o-mini exhibit wrong reasoning hallucinations leading to incorrect diagnoses and BioReason hallucinates to partially correct diagnoses.

Figure 7: Qualitative comparison after gene-name anonymization. Despite the removal of explicit gene identifiers, GenoMorph correctly infers the disease by reasoning from genomic sequence information, whereas GPT-4o-mini and BioReason fail under the anonymized setting.

Disease	Misses	Observation
Creutzfeldt–Jakob disease	7	Predicted broader category (Prion disease)
Sphingolipidoses	1	Predicted specific subtype
Hepatocellular carcinoma	2	Genuine prediction errors
Pancreatic ductal adenocarcinoma	1	Genuine prediction error
Colorectal cancer	1	Genuine prediction error
Non-small cell lung cancer	1	Genuine prediction error

Table 8: Manual analysis of the remaining prediction errors produced by GenoMorph with self-adaptive RSFT.

Explore similar work

When Do Biological Reasoning Models Use Their Biological Inputs?
Oct 1, 2026Ada Fang, Nikitha Thoduguli, Lukas Fesser +3Evidence Conflict Foundation Model
Harvard University · Massachusetts Institute of Technology
HypoKG: Evidence-Disciplined Biomedical Hypothesis Generation Beyond Endpoint Knowledge
Sep 14, 2026Dominic Okonkwo, Adetayo Okunoye, Ismailcem Budak ArpinarBiomedical Knowledge Graph Hypothesis-Driven Model Expansion
School of Computing University of Georgia
Tool-Augmented On-Policy Distillation for LLM Domain Adaptation in Sequence-Based Omics Tasks
Sep 20, 2026Jie Ying, Zhefan Wang, Zihong Chen +11Multi-Omic Integration Reasoning Benchmark
Shanghai Artificial Intelligence Laboratory · Yazhouwan National Laboratory · Shanghai Innovation Institute +2

Oct 1, 2026, cs.LGJ/K move · Enter open · S save

When Do Biological Reasoning Models Use Their Biological Inputs?

Ada Fang, Nikitha Thoduguli, Lukas Fesser, Hanlin Zhang+2

Harvard University · Massachusetts Institute of Technology

Evidence Conflict Foundation Model Intermediate Checkpoints

Open paper AskPDF arXiv

Biological reasoning models use post-training to connect LLMs to biological foundation model representations and biological text. Their benchmark accuracy is taken as evidence that LLMs reason over these inputs. We test this assumption in six biological reasoning models across DNA, protein, and single-cell tasks. We perturb one biological input while holding the query and other inputs fixed, construct evidence conflicts that pair the foundation model representation of one genome, protein, or cell with the text of another, fit linear probes to the representations the language model receives, and analyze reasoning traces against the biological inputs. Evo2 and ESM3 contribute little to BioReason and BioReason-Pro performance on the evaluated tasks. Shuffling the DNA sequence barely changes BioReason disease prediction accuracy, and in evidence conflicts the two models follow the text in 97.9% and 99.7% of cases. Linear probes trained on the Evo2 and ESM3 representations predict the task targets, so these foundation models encode information relevant to the task, but provide limited overall performance improvement to BioReason and BioReason-Pro. In contrast, foundation model inputs contribute to ChatNT, Prot2Text-V2, and CellWhisperer performance, and differentially expressed genes in the gene sentence contribute to Cell2Sentence-Scale performance. Across SFT and RL checkpoints of BioReason-Pro and 42 BioReason checkpoints, increases in accuracy do not imply greater performance contributions from biological inputs. BioReason traces misstate nucleotide changes, while BioReason-Pro traces describe functions omitted from final predictions under evidence conflicts. We find that current post-training strategies do not ensure that foundation model representations contribute to task performance.

Abstract

Explore similar work

When Do Biological Reasoning Models Use Their Biological Inputs?

HypoKG: Evidence-Disciplined Biomedical Hypothesis Generation Beyond Endpoint Knowledge

Tool-Augmented On-Policy Distillation for LLM Domain Adaptation in Sequence-Based Omics Tasks