PreAcrs: a machine learning framework for identifying anti-CRISPR proteins

BMC Bioinformatics

Table 5 Features: the sequence and structural features calculated and their dimensionalities

Feature type	Feature cluster	Dimensions	Reduced-dimensions
Sequence	AAC	20	20
	PAAC	23	23
	CKSAAP	2400	200
	DDE	400	200
	DPC	400	200
Evolutionary	PSSM-composition	400	200
	DPC-PSSM	400	200
	PSSM-AC	200	200
	RPSSM	110	110
	PSSM-SMTH	1000	200
Pre-trained	BiLSTM	3605	200
	LM	533	200
	SSA	121	121
	TAPE-BERT	768	200
	UniRep	1900	200
	W2V	300	200
	esm	1280	200
	ProtTrans	1024	200
Total		14,884	3074

ISSN: 1471-2105