Back

Incorporating LLM-Derived Information into Hypothesis Testing for Genomics Applications

Bryan, J. G.; Niu, H.; Li, D.

2025-05-07 bioinformatics
10.1101/2025.04.30.651464 bioRxiv
Show abstract

We propose strategies for incorporating the information in large language models (LLMs) into statistical hypothesis tests in genomics studies. Using gene embeddings derived from text inputs to OpenAIs GPT-3.5 model, we show that biological signals in a variety of genomics datasets reside near the principal subspace spanned by the embeddings. We then use a frequentist and Bayesian (FAB) framework to propose several hypothesis tests that are either optimal or approximately optimal with respect to prior information based on the gene embedding subspace. In four real-world genomics examples, the FAB tests guided by the LLM-derived information achieve more power than classical counterparts.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

1
Bioinformatics
1204 papers in training set
Top 0.8%
22.9%
2
Biostatistics
24 papers in training set
Top 0.1%
15.5%
3
Journal of Computational Biology
48 papers in training set
Top 0.1%
6.4%
4
BMC Bioinformatics
457 papers in training set
Top 2%
4.4%
5
Biometrics
23 papers in training set
Top 0.1%
4.4%
50% of probability mass above
6
The Annals of Applied Statistics
19 papers in training set
Top 0.1%
4.1%
7
Briefings in Bioinformatics
354 papers in training set
Top 2%
3.6%
8
PLOS Computational Biology
1863 papers in training set
Top 9%
3.3%
9
Bioinformatics Advances
203 papers in training set
Top 2%
3.3%
10
Cell Systems
201 papers in training set
Top 2%
2.7%
11
Frontiers in Genetics
230 papers in training set
Top 2%
2.5%
12
Statistics in Medicine
40 papers in training set
Top 0.3%
2.0%
13
BMC Genomics
406 papers in training set
Top 4%
1.8%
14
GENETICS
483 papers in training set
Top 2%
1.8%
15
Genome Biology
637 papers in training set
Top 5%
1.8%
16
Patterns
78 papers in training set
Top 2%
1.4%
17
Genetic Epidemiology
55 papers in training set
Top 0.6%
1.1%
18
BioData Mining
22 papers in training set
Top 0.5%
1.1%
19
IEEE/ACM Transactions on Computational Biology and Bioinformatics
38 papers in training set
Top 0.9%
1.1%
20
NAR Genomics and Bioinformatics
242 papers in training set
Top 4%
0.9%
21
Nature Communications
5641 papers in training set
Top 58%
0.6%
22
G3: Genes, Genomes, Genetics
252 papers in training set
Top 5%
0.6%
23
Scientific Reports
3612 papers in training set
Top 77%
0.6%