FlowingFlowing
← Back to all development cases
This case was judged by an anonymized AI reviewer and has not been independently reviewed by people; inspect the draft, cards, and both outputs to assess it.
AutocompleteWithout cards better150 / 280 · beu-6e9236ff27d38960

Large Language Models in Bioinformatics: A Survey

Quantitative biology · 2503.04490v3

FLOWING EVIDENCE BENCHMARK

How do we compare writing with and without cards?

At the same writing position, the same model produces one first result with cards and one without. All 280 pairs enter the overall anonymized review, which assesses whether a decision not to continue fits the draft.

Overall: 138 with-card wins, 26 ties, 107 no-card wins, 5 pairs where neither was usable, and 4 inconclusive judgments. Both sides wrote text in 218 pairs; 48 had a deliberate stop and 14 included an invalid output.

SAME DRAFT · WITH OR WITHOUT CARDS

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED VERSIONS

SHARED BY BOTH

Draft, writing task, and generation model

A · With cards

Relevant literature cards supplied

B · Without cards

No literature cards supplied

Generation model

Same model and version

A → first output

B → first output

Anonymized AI judge

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

Review again with the answer order swapped

HOW DOES BLIND JUDGING WORK?

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

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 positions and cards selected?

We selected 40 writing positions in each of seven fields, for 280 comparisons. The position and relevant literature cards were set before each comparison, so both versions began from the same draft.

The main result retains every frozen position, including places where cards might not add a useful next move. This avoids selecting cases based on their outputs, but 49% describes the whole cohort, not a win rate where evidence clearly had an opportunity to help. This cohort did not record that opportunity label for each position before generation, so those two groups cannot yet be separated reliably.

CARD USE · 140/280

Reported card use in continuations with text

The model reports whether a card affected its continuation. In 140 of 280 tests it produced continuation text and reported card use. The previous count of 150 included 10 cases without continuation text and has been corrected.

COMPARISON · 70 / 7 / 30

Which written version was better?

Of the 107 reported-use cases where both versions produced text, the blinded AI judge preferred the with-card version 70 times, found 7 about the same, and preferred the without-card version 30 times.

How are the two table percentages calculated?

ADOPTION COVERAGE

140reported use with text
÷
280all positions
=
50%

WITH-CARD VERSION PREFERRED

70with cards better
÷
107comparable tests
=
65%

The percentages use different groups: 50% counts continuations with text reporting card use among all 280 tests; 65% covers the 107 such cases with text on both sides. The generating model reports card use, while an anonymized AI judge compares writing quality.

Card use is self-reported by the generator. The anonymized AI judge sees the draft and source snippets, so source-aware preference remains possible; there was no independent human review. The overall result describes these 280 frozen positions, not a win probability for any writing task.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

Large Language Models in Bioinformatics: A Survey

1 Introduction

…l data. This survey provides a systematic review of recent advancements, focusing on genomic sequence modeling, RNA structure prediction, protein function inference, and single-cell transcriptomics. Meanwhile, we also discuss several key challenges, including data scarcity, computational complexity, and cross-omics integration, and explore future directions such as multimodal learning, hybrid AI models, and clinical applications. By offering a comprehensive perspective, this paper underscores the transformative potential of LLMs in driving innovations in bioinformatics and precision medicine.

↓ 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 positionOverall AI review: Without cards better

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.

Read overall AI review rationale

A = without cards · B = with cards

The draft is a complete abstract/conclusion paragraph that already achieves closure. It systematically covers: scope (genomic sequence modeling, RNA structure prediction, protein function inference, single-cell transcriptomics), challenges (data scarcity, computational complexity, cross-omics integration), future directions (multimodal learning, hybrid AI models, clinical applications), and a concluding significance statement. Output A correctly abstains with 'complete' because continuing would be redundant—the paragraph already has a clear beginning, middle, and end, with no hanging argument requiring resolution. Output B adds a new thought about foundation models handling unlabeled data, which, while topically related, introduces a new conceptual thread (foundation models vs. LLMs) that wasn't in the original paragraph's framing. The draft specifically discusses 'LLMs' throughout, not 'foundation models' broadly, and B's pivot to FMs creates a slight terminological mismatch. Moreover, B's content about unlabeled data and downstream tasks partially overlaps with but doesn't advance the already-completed argument. The paper has already stated its comprehensive perspective and transformative potential; adding more would be speculative extension rather than necessary continuation. A's abstention is the more precise judgment of what this draft needs.

Evidence

Same model and position · retrieved passages supplied · first output

Not preferred in this comparison

The rapid growth of biological data, together with the scarcity of experimentally determined labels, has made large-scale unlabeled datasets a defining feature of modern bioinformatics. Foundation models are particularly adept at handling such data, which has allowed them to achieve notable results across a range of downstream biological tasks.

Passages supplied to the Evidence version

Card 1 · Progress and Opportunities of Foundation Models in Bioinformatics

Bioinformatics has witnessed a paradigm shift with the increasing integration of artificial intelligence (AI), particularly through the adoption of foundation models (FMs). These AI techniques have rapidly advanced, addressing historical challenges in bioinformatics such as the scarcity of annotated…

Read full passage excerpt

Bioinformatics has witnessed a paradigm shift with the increasing integration of artificial intelligence (AI), particularly through the adoption of foundation models (FMs). These AI techniques have rapidly advanced, addressing historical challenges in bioinformatics such as the scarcity of annotated data and the presence of data noise. FMs are particularly adept at handling large-scale, unlabeled data, a common scenario in biological contexts due to the time-consuming and costly nature of experimentally determining labeled data. This characteristic has allowed FMs to excel and achieve notable results in various downstream validation tasks, demonstrating their ability to represent diverse biological entities effectively. Undoubtedly, FMs have ushered in a new era in computational biology, especially in the realm of deep learning. The primary goal of this survey is to conduct a systematic investigation and summary of FMs in bioinformatics, tracing their evolution, current research status …

Card 2 · RadioPathomics: Multimodal Learning in Non-Small Cell Lung Cancer for Adaptive Radiotherapy

Current practice in cancer treatment collects multimodal data, such as radiology images, histopathology slides, genomics and clinical data. The importance of these data sources taken individually has fostered the recent rise of radiomics and pathomics, i.e., the extraction of quantitative features f…

Read full passage excerpt

Current practice in cancer treatment collects multimodal data, such as radiology images, histopathology slides, genomics and clinical data. The importance of these data sources taken individually has fostered the recent rise of radiomics and pathomics, i.e., the extraction of quantitative features from radiology and histopathology images collected to predict clinical outcomes or guide clinical decisions using artificial intelligence algorithms. Nevertheless, how to combine them into a single multimodal framework is still an open issue. In this work, we develop a multimodal late fusion approach that combines hand-crafted features computed from radiomics, pathomics and clinical data to predict radiotherapy treatment outcomes for non-small-cell lung cancer patients. Within this context, we investigate eight different late fusion rules and two patient-wise aggregation rules leveraging the richness of information given by CT images, whole-slide scans and clinical data. The experiments in le …

Card 3 · TRONCO: an R package for the inference of cancer progression models from heterogeneous genomic data

MOTIVATION: We introduce TRanslational ONCOlogy (TRONCO), an open-source R package that implements the state-of-the-art algorithms for the inference of cancer progression models from (epi)genomic mutational profiles. TRONCO can be used to extract population-level models describing the trends of accu…

Read full passage excerpt

MOTIVATION: We introduce TRanslational ONCOlogy (TRONCO), an open-source R package that implements the state-of-the-art algorithms for the inference of cancer progression models from (epi)genomic mutational profiles. TRONCO can be used to extract population-level models describing the trends of accumulation of alterations in a cohort of cross-sectional samples, e.g. retrieved from publicly available databases, and individual-level models that reveal the clonal evolutionary history in single cancer patients, when multiple samples, e.g. multiple biopsies or single-cell sequencing data, are available. The resulting models can provide key hints for uncovering the evolutionary trajectories of cancer, especially for precision medicine or personalized therapy. AVAILABILITY AND IMPLEMENTATION: TRONCO is released under the GPL license, is hosted at http://bimib.disco.unimib.it/ (Software section) and archived also at bioconductor.org. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: S …

Model-reported card use

Card 1 · fact

Affected output: The rapid growth of biological data, together with the scarcity of experimentally determined labels, has made large-scale unlabeled datasets a defining feature of modern bioinformatics.

Model explanation: The row supplied the premise that unlabeled data dominate biological contexts because experimental labeling is time-consuming and costly, which grounds the opening sentence of the new paragraph.

Card 1 · terminology

Affected output: Foundation models are particularly adept at handling such data

Model explanation: The row provided the field-native phrasing 'particularly adept at handling' for describing foundation models' relation to unlabeled data.

No Evidence

Same model and position · no retrieved passages · first output

Preferred by blind review

No continuation text was produced.

Model action：complete

Earlier card-use comparison

The earlier diagnostic compared only cases with reported card use and text on both sides; its scope differs from the overall AI review above.

The with-card arm produced text and reported card use, but the no-card arm produced no text. This counts as reported use, without entering the two-text preference comparison.

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.