Back

Generalist large language models complement tailor-made predictors for tumor genomics interpretation

Yu, J.; Darmofal, M.; Waters, M.; Choy, J.; Tran, T. N.; Fu, C.; Morales, L.; U, K.; Levine, R. L.; Schultz, N.; Berger, M. F.; Morris, Q.; Jee, J.

2026-05-22 genomics
10.64898/2026.05.21.726957 bioRxiv
Show abstract

General-purpose large language models (LLMs) are trained on large corpora to acquire broad knowledge, but whether LLMs can replace, or augment, task-specific models is unclear. We evaluated LLMs on three real-world, clinically important tumor genomic interpretation tasks, in order of increasing difficulty: (i) distinguishing tumor from non-tumor mutations (n=34,415 variants), (ii) distinguishing driver from passenger mutations (n=13,469 variants), and (iii) inferring cancer type from tumor sequencing reports across multiple assays and institutions (n=102,791 samples). The best general-purpose LLMs performed as well as the benchmark tailor-made predictor for task (i). Ensembling tailor-made models with zero-shot LLMs improved their performance for tasks (i) and (ii). For task (iii), LLMs outperformed or supplemented tailor-made models on out-of-distribution data. Without fine-tuning, current LLMs already can be useful in clinical genomic interpretation by adding complementary expertise to tailor-made, state-of-the-art predictors.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
Genome Medicine
154 papers in training set
Top 0.1%
28.3%
2
Nature Machine Intelligence
61 papers in training set
Top 0.4%
6.5%
3
Genome Biology
555 papers in training set
Top 1%
6.5%
4
Nature Communications
4913 papers in training set
Top 35%
4.4%
5
Nature Biotechnology
147 papers in training set
Top 2%
3.7%
6
Nature
575 papers in training set
Top 8%
3.3%
50% of probability mass above
7
Nature Medicine
117 papers in training set
Top 1%
3.1%
8
Nature Genetics
240 papers in training set
Top 3%
3.1%
9
Bioinformatics
1061 papers in training set
Top 6%
2.9%
10
Nature Methods
336 papers in training set
Top 3%
2.8%
11
Genome Research
409 papers in training set
Top 2%
2.1%
12
Cell Genomics
162 papers in training set
Top 2%
2.1%
13
Briefings in Bioinformatics
326 papers in training set
Top 3%
1.9%
14
PLOS Computational Biology
1633 papers in training set
Top 14%
1.9%
15
The American Journal of Human Genetics
206 papers in training set
Top 2%
1.9%
16
Scientific Reports
3102 papers in training set
Top 56%
1.7%
17
Nucleic Acids Research
1128 papers in training set
Top 10%
1.7%
18
Nature Biomedical Engineering
42 papers in training set
Top 1%
1.3%
19
Science
429 papers in training set
Top 17%
1.3%
20
Genetics in Medicine
69 papers in training set
Top 0.9%
0.8%
21
BMC Bioinformatics
383 papers in training set
Top 6%
0.8%
22
Frontiers in Genetics
197 papers in training set
Top 9%
0.8%
23
PLOS ONE
4510 papers in training set
Top 70%
0.7%
24
BMC Genomics
328 papers in training set
Top 6%
0.7%
25
Proceedings of the National Academy of Sciences
2130 papers in training set
Top 47%
0.7%
26
Nature Human Behaviour
85 papers in training set
Top 5%
0.7%
27
Annals of Oncology
13 papers in training set
Top 1%
0.5%
28
Molecular Systems Biology
142 papers in training set
Top 2%
0.5%
29
European Journal of Human Genetics
49 papers in training set
Top 2%
0.5%
30
Communications Biology
886 papers in training set
Top 31%
0.5%