notesum.ai

Published at November 25

Deciphering genomic codes using advanced NLP techniques: a scoping review

q-bio.GN

cs.AI

Released Date: November 25, 2024

Authors: Shuyan Cheng¹, Yishu Wei¹, Yiliang Zhou¹, Zihan Xu¹, Drew N Wright², Jinze Liu³, Yifan Peng¹

Aff.: ¹Department of Population Health Sciences, Weill Cornell Medicine; ²Samuel J. Wood Library & C.V. Starr Biomedical Information Center, Weill Cornell Medicine; ³School of Public Health, Virginia Commonwealth University

Arxiv: http://arxiv.org/abs/2411.16084v1

Model	Accuracy	F1	MCC	ROCAUC	Specificity	Precision	Recall
BERT-5mC [18]	0.933	-	0.656	0.966	0.938	-	0.872
DNABERT [12]^a	0.965	0.965	-	0.930	-	-	-
SETOMIC [34]	0.950	0.921	-	0.997	-	0.945	-
SETQUENCE [34]	0.475	0.359	-	0.910	-	0.375	-
BERT-CNN [13]	0.756	-	0.514	-	0.712	-	0.800
TFBERT [15]	0.880	0.880	0.762	0.947	-	0.882	0.880
IGnet [35]	0.838	0.824	-	0.924	-	0.875	0.778
MuLan-Methyl [25]	0.948	0.950	-	0.968	-	-	0.979
moDNA [22]	0.862	0.862	0.725	0.935	-	0.863	0.862
DistilBERT+CRF+Attention Mask [11]	0.965	0.735	-	-	0.959	0.691	0.852
BERT+CRF (with/without) [23]	0.973	0.834	-	-	0.962	0.780	0.897
BERT-Promoter [14]	0.855	-	-	-	0.866	-	0.843
DeepViFi (pipeline) [16]	0.960	-	-	0.94	-	0.996	1.000
GENEMASK-based [17]	0.898	-	-	0.962	-	-	-
MSCAN [26]	0.957	0.713	0.710	0.937	0.994	0.905	-
MTTLm6 [27]	0.699	0.713	0.399	0.771	0.649	0.681	-
BioDGW-CMI [29]^a	0.885	0.885	-	0.948	-	0.885	0.885
BCMCMI [28]	0.832	0.836	0.667	0.904	-	0.808	0.868
MiTDS [30]	0.770	0.810	-	-	-	-	0.960