FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteEvidence version preferred93 / 162 · 5c824a3a2b6da84c

Large Language Models in Bioinformatics: A Survey

Quantitative biology · 2503.04490v3

FLOWING EVIDENCE BENCHMARK

How do we tell whether Evidence helps?

At the same writing position, with the same model and task, how does supplying retrieved paper passages change the first output? We compare matched versions and retain ties, unusable outputs and incomplete reviews.

SAME DRAFT · EVIDENCE ON OR OFF

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED INPUTS

SHARED BY BOTH

Manuscript context, model, task and prompt

A · With Evidence

Retrieved paper passages supplied

B · Without Evidence

No retrieved passages supplied

LLM

Same model and version

A → first output

B → first output

Blind judge agent

First outputs are anonymized as X and Y

Continuation review: accuracy, fit to the writing task and usability
Returns: X preferred / tie / Y preferred / both unusable

Order check: X / Y → Y / X

HOW DOES BLIND JUDGING WORK?

① Anonymize both outputs
The judge sees the same draft and both first outputs without knowing which received Evidence.

② Compare and swap order
The judge applies task-specific criteria in X/Y and then Y/X order.

③ Review disagreements
A third pass resolves disagreements. Incomplete reviews remain in the denominator.

How are autocomplete positions stratified?

Before seeing generation outcomes, we check whether a retrieved passage contains a specific proposition that directly supports the next writing move. Those positions appear in the left opportunity group; the rest are ordinary positions on the right. We select a balanced sample from admitted papers in each field. The 50/50 split is experimental, not a measure of how often either type occurs in writing.

AUTOCOMPLETE · 110

Biology, statistics and astrophysics

55 positions on each side; statistics uses the ten-paper rerun.

AUTOCOMPLETE · 52

Psychology and climate science

26 positions on each side; psychology includes nine papers and climate science four.

How are the table percentages calculated?

Across five fields there are 81 source-opportunity positions. The original blind review preferred Evidence in 51; a task check moved one empty Evidence continuation to both unusable, leaving 50 in public counts. In the ordinary group, another pair of empty outputs moved from no winner to both unusable. Original verdicts remain visible on case pages.

50Evidence version preferred
÷
81All positions in this group
=
62%Evidence preference in this group

Source contribution is a separate review: 16 of 21 Evidence wins entered into source review directly used retrieved papers; another 29 wins await review.

These are development-stage model judgments pending independent human review. They are not formal Benchmark conclusions and do not, on their own, establish causality.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

Large Language Models in Bioinformatics: A Survey

7 Conclusions and Future Directions > 7.1 Challenges and Limitations > 7.1.2 Computational Complexity and Model Efficiency

In conclusion, this paper comprehensively examined the applications of LLMs across DNA, RNA, protein, and single-cell data analysis, highlighting key research contributions and emerging methodologies. Despite advances, LLM applications in bioinformatics remain evolving, requiring key challenges to address for full potential. Therefore, we here discuss the current limitations and outline promising future directions for advancing LLM-driven bioinformatics research.

… large-scale, high-quality biological datasets for effective training, yet annotated genomic, transcriptomic, and proteomic data remain limited Lu et al. 2024 . Unlike natural language corpora, which are abundant and diverse, biological datasets are often noisy, incomplete, or biased toward well-studied species and diseases. Consequently, model generalizability suffers, leading to biased predictions which may not hold across diverse biological contexts. Additionally, batch effects and experimental noise complicate the development of robust foundation models for bioinformatics Yu et al. 2024 .

↓ Next paragraph was generated here

The highlighted text is near the writing position. This is a layout of the evaluation input, not a PDF screenshot.

A matched comparison at the same writing positionJudgment: Evidence version preferred

Both arms use the same model, task and manuscript context. Only the retrieved passages differ. The first recorded output from each arm is shown below.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

Beyond data limitations, the computational demands of LLMs pose a further barrier to their routine adoption in bioinformatics. Training and deploying large models require substantial GPU memory, storage, and energy, which are often unavailable in typical academic or clinical settings. This resource gap constrains who can develop and apply these models, reinforcing the concentration of LLM research in well-resourced institutions.

Passages supplied to the Evidence version

Heterogeneous large datasets integration using Bayesian factor regression ↗

Equation ( 1 ) regresses the observed data X X on known covariates and on a latent factor structure. In particular, it allows additive batch effects to be accounted for by incorporating the variables recording the batches into v i \texttt{v}_{i} . However, in practice one often observes more complex…

Read full passage excerpt

Equation ( 1 ) regresses the observed data X X on known covariates and on a latent factor structure. In particular, it allows additive batch effects to be accounted for by incorporating the variables recording the batches into v i \texttt{v}_{i} . However, in practice one often observes more complex batch effects; specifically in bioinformatics it is common to observe multiplicative effects on the variance ( Johnson et al., 2007 ) . We will later describe an example of this, shown in Figure 5 . Such artefacts cannot be captured by ( 1 ) given that Σ \Sigma is assumed constant across all individuals.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

The product produced no insertable continuation text.

The model explicitly returned a no-completion marker.

Judgment and task check

The blinded model review preferred the Evidence version. Its original rationale appears below.

Read the original model judgment

In this rationale, Output A = No Evidence; Output B = Evidence.

Output A is empty when the task clearly requires content—a coherent academic paragraph continuing the manuscript. The draft sets up a discussion of 'current limitations and promising future directions' for LLM-driven bioinformatics, and the preceding paragraph specifically discusses data limitations (scarcity, noise, bias, batch effects). Output A's empty text cannot be inserted or used. Output B provides a coherent continuation that transitions from data limitations to computational demands as another barrier, maintaining the manuscript's scope and academic tone. The claims in B about GPU memory, storage, energy requirements, and resource concentration in well-resourced institutions are general, plausible statements about LLM computational demands that do not require source support—they are established background knowledge about LLMs, not specific factual claims needing citation. No generated citation markers are present. The paragraph logically extends the limitations discussion without inventing this paper's own methods or results.

The passages are input material; their presence does not mean an output used them or that they support every claim. Draft and source passages are excerpted; no simulated scores or PDF appear here.