MicroRNA based Pan-Cancer Diagnosis and Treatment Recommendation

BMC Bioinformatics

Table 1 Distribution of the cancer and normal samples in the dataset used to build the predictive model (TCGA cancer classifier)

Organ/System	Cancer	Cancer acronym	Normal samples	Cancer samples
Thymus	Thymoma	THYM	2	124
Lung	Lung Squamous Cell Carcinoma	LUSC	45	342
Pancreas	Pancreatic Adenocarcinoma	PAAD	4	179
GI tract	Cholangiocarcinoma	CHOL	9	36
	Esophageal Carcinoma	ESCA	13	185
	Stomach Adenocarcinoma	STAD	41	395
Liver	Liver Hepatocellular Carcinoma	LIHC	50	374
Thyroid	Thyroid Carcinoma	THCA	59	510
Adipose	Adrenocortical carcinoma	ACC	0	80
Lymph	Diffuse Large B-cell Lymphoma	DLBC	0	47
Heart	Mesothelioma	MESO	0	87
Reproductive	Cervical Squamous Cell and Endocervical Adenocarcinoma	CESC	3	309
	Ovarian Serous Cystadenocarcinoma	OV	0	461
	Testicular Germ Cell Tumors	TGCT	0	156
Urinary	Uterine Carcinosarcoma	UCS	0	56
Kidney	Kidney Chromophobe	KICH	25	66
Kidney	Kidney Renal Papillary cell carcinoma	KIRP	34	292
Brain	Brain Lower Grade Glioma	LGG	0	526
Peripheral Nervous System	Pheochromocytoma and Paraganglioma	PCPG	3	184
Epidermis	Skin Cutaneous Melanoma	SKCM	2	450
Epidermis	Uveal Melanoma	UVM	0	80

Note that not all the cancer types have normal samples. Even though the TCGA dataset has about 33 cancer types, many cancer types were removed due to lack of data (less than 5 samples) and in the end 21 cancer types as listed in this table were used for the classification

ISSN: 1471-2105