S1. Number of filtered sequence reads per specimen*.

Cross-sections	Total (N=428 specimens)			Cases (N=143 specimens)**			Controls (N=285 specimens)**
	n	Reads ( )		n	Reads ( )		n	Reads ( )
*, Quality filtering on the raw reads was performed under specific filtering conditions to obtain the high-quality clean reads according to the Cutadapt (Version 1.9.1, http://cutadapt.readthedocs.io/en/stable/) quality control process. The UCHIME algorithm (UCHIME Algorithm, http://www.drive5.com/usearch/manual/uchime_algo.html) was used to detect chimera sequences, and then the chimera sequences were removed. Finally, clean reads were obtained. **, Cases were subjects with esophageal lesions of severe dysplasia and above including severe squamous dysplasia, carcinoma in situ, and esophageal squamous cell carcinoma. Controls were subjects without SDA lesions.
1	243	76,104±10,786		81	76,285±10,730		162	76,013±10,846
2	109	78,327±9,729		38	78,198±8,970		71	78,396±10,174
3	20	77,439±14,943		7	84,419±21,085		13	73,680±9,356
4	12	79,233±7,502		4	83,016±5,076		8	77,342±8,071
5	12	77,158±8,048		4	82,153±3,490		8	74,660±8,670
6	10	77,583±9,607		2	71,930±11,659		8	78,996±9,372
7	7	74,729±5,573		2	76,054±5,862		5	74,200±6,064
8	15	77,513±8,925		5	73,569±9,946		10	79,486±8,184
Total	428	76,911±10,444		143	77,385±10,733		285	76,673±10,306