Use of Biochemical Tests and Machine Learning in the Search for Potential Diagnostic Biomarkers of COVID-19, HIV/AIDS, and Pulmonary Tuberculosis

Cobre, Alexandre F.; Morais, Amiel A.; Selege, Fosfato; Stremel, Dile P.; Wiens, Astrid; Ferreira, Luana M.; Tonine, Fernanda S.; Pontarolo, Roberto

doi:10.21577/0103-5053.20240020

Acessibilidade / Reportar erro

Brasil

Journal of the Brazilian Chemical Society

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Article • J. Braz. Chem. Soc. 35 (7) • 2024 • https://doi.org/10.21577/0103-5053.20240020 copy

Use of Biochemical Tests and Machine Learning in the Search for Potential Diagnostic Biomarkers of COVID-19, HIV/AIDS, and Pulmonary Tuberculosis

Authorship SCIMAGO INSTITUTIONS RANKINGS

This study aims to develop, validate, and evaluate machine learning algorithms for predicting the diagnosis of coronavirus disease (COVID-19), human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS), pulmonary tuberculosis (TB), and HIV/TB co-infection. We also investigated potential biomarkers associated with the diagnosis. Data from biochemical and hematological tests of infected and controls were collected in a single general hospital, totalizing 6,418 patients. The discriminant analysis by partial least squares (PLS-DA) model had the highest performance in predicting the diagnosis of COVID-19, HIV/AIDS, TB, and HIV/TB co-infection with an accuracy of 94, 97, 95, and 96%, respectively. The biomarkers calcium, lactate dehydrogenase, red blood cells (RBC), white blood cells, neutrophils, basophils, eosinophils, hemoglobin, and hematocrit were associated with COVID-19. HIV infection was associated with mean corpuscular volume, platelets, neutrophils, and mean platelet volume. Red blood cell distribution width and urea were associated with infection by Mycobacterium tuberculosis. The following biomarkers were associated with HIV/TB co-infection: lymphocytes, RBC, hematocrit, hemoglobin, aspartate transaminase, alanine transaminase, and glycemia. The PLS-DA model can optimize COVID-19, HIV/AIDS, TB, and HIV/TB co-infection diagnostics. Some biomarkers were potential diagnostic indicators and could be evaluated during the screening of these diseases.

Keywords:
COVID-19; HIV/AIDS; tuberculosis; co-infection; diagnosis; machine learning

Test name	Measurement unit
White blood cells	(× 10⁹ L^-1)
Red blood cells	(× 10¹² L^-1)
Hemoglobin	(g dL^-1)
Hematocrit	%
Mean corpuscular volume	fL
Mean corpuscular hemoglobin	(g cell^-1)
Mean corpuscular hemoglobin	(g Hb dL^-1)
Platelets count	(× 10⁹ L^-1)
Platelets count	%
Platelets count	(× 10⁹ L^-1)
Neutrophils count	%
Neutrophils count	(× 10⁹ L^-1)
Monocytes count	%
Monocytes count	(× 10⁹ L^-1)
Creatinine Kinase	(U L^-1)
Urea	(mg dL^-1)
Aspartate aminotransferase	(U L^-1)
Alanine aminotransferase	(U L^-1)
Glucose	(mg dL^-1)
Calcium	(mmol L^-1)
Sodium	(mmol L^-1)
Potassium	(mmol L^-1)

Dataset			Training (70%)		Testing
Dataset	Positive	Negative (control)	Positive	Negative	Positive	Negative
COVID-19	816	920	571	644	245	276
HIV/AIDS	49	4520	34	3164	15	1356
TB	113	4520	79	3164	34	1356
HIV/TB	80	4520	56	3164	24	1356

Variable	COVID-19		HIV/AIDS		Tuberculosis		HIV/TB		Control
Variable	Median	IQR^a	Median	IQR^a	Median	IQR^a	Median	IQR^a	Median	IQR^a	P^b
Age / years	62.00	51.00-75.00	38.00	32.00-43.00	32.00	24.00-40.00	33.50	28.25-42.75	64.00	45.00-78.00	0.00
White blood cells / (×10⁹ L-¹)	6.58	5.01-9.58	4.50	3.70-5.60	7.30	5.15-9.75	7.15	4.83-9.98	8.80	6.75-11.31	0.00
Red blood cells / (×10¹² L-¹)	4.67	4.27-5.07	3.93	3.46-4.75	4.48	3.99-4.94	3.73	3.27-4.33	4.50	4.03-4.89	0.00
Hemoglobin / (g dL^-1)	13.70	12.31-14.74	12.40	11.25-13.85	11.20	9.85-13.15	10.00	8.53-11.43	13.10	11.53-14.30	0.00
Hematocrit / %	40.50	37.00-43.55	38.10	34.65-40.90	35.00	31.35-39.60	30.30	26.55-34.40	39.03	35.30-42.20	0.00
Mean corpuscular volume / fL	87.10	84.00-90.10	98.90	76.95-107.85	73.00	-70.30-83.95	71.05	-73.88-87.78	87.70	84.26-91.24	0.00
Mean corpuscular hemoglobin / (g cell^-1)	29.40	28.21-30.50	31.20	27.90-34.45	25.20	-19.30-27.75	26.50	-17.75-29.05	29.50	28.10-30.70	0.00
Mean corpuscular hemoglobin / (g Hb dL^-1)	33.70	32.90-34.35	31.90	29.70-34.55	31.90	30.00-33.25	32.05	28.98-33.80	33.50	32.56-34.20	0.00
Platelets count / (×10⁹ L-¹)	209.00	162.00-268.75	212.00	159.50-268.50	319.50	226.50-396.00	305.50	226.00-389.75	234.67	187.00 -284.00	0.00
Platelets count / %	15.65	10.16-22.69	41.10	27.85-51.70	27.20	21.00-37.90	24.20	14.80-37.75	18.40	14.18-22.60	0.00
Platelets count / (×10⁹ L-¹)	1.00	0.71-1.30	2.20	1.75-2.95	2.00	1.50-2.65	1.60	1.10-2.38	1.50	1.20-1.75	0.00
Neutrophils count / %	75.60	66.83-83.50	31.80	20.65-47.35	58.50	52.45-66.60	57.35	47.15-76.75	70.40	66.40-76.10	0.00
Neutrophils count / (×10⁹ L-¹)	4.80	3.50-7.25	1.80	1.34-2.35	4.40	3.06-6.70	4.30	3.03-6.28	5.95	5.00-7.10	0.00
Monocytes count / %	7.30	5.20-9.50	10.70	8.00-15.15	10.30	8.30-12.80	9.15	7.08-11.88	7.78	6.80-8.70	0.00
Monocytes count / (×10⁹ L-¹)	0.50	0.30-0.60	0.50	0.35-0.68	0.80	0.60-1.10	0.80	0.50-0.90	0.60	0.56-0.75	0.00
Creatinine kinase / (U L^-1)	100.00	100.00-100.00	83.85	76.15-91.35	61.83	61.83-61.83	63.41	63.41-63.41	78.00	78.00-78.00	0.00
Urea / (mg dL^-1)	34.83	30.00-42.00	3.55	3.10-3.80	3.40	3.40-3.40	4.74	4.74-4.74	35.00	33.00-38.30	0.00
Aspartate aminotransferase / (U L^-1)	43.00	30.00-63.00	31.00	26.50-34.50	29.30	29.30-29.30	23.40	23.40-23.40	26.00	21.00-33.00	0.00
Alanine aminotransferase / (U L^-1)	33.85	23.00-54.46	31.00	27.00-38.35	21.00	21.00-21.00	25.20	25.20-25.20	22.00	16.31-31.00	0.00
Glucose / (mg dL^-1)	107.00	95.00-129.00	4.41	4.41-4.41	4.70	4.70-4.70	4.82	4.82-4.82	100.75	89.00-118.00	0.00

Class	Model	TP	FN	TN	FP	Sensitivity	Specificity	Accuracy	Precision	F1 score	MCC
COVID-19	PLS-DA	226	19	265	11	0.92	0.96	0.94	0.95	0.94	0.88
	ANN	224	21	248	28	0.91	0.90	0.91	0.89	0.90	0.81
	KNN	219	26	259	17	0.89	0.94	0.92	0.93	0.91	0.83
	SVM	205	40	259	17	0.84	0.94	0.89	0.92	0.88	0.78
	SIMCA	238	7	244	32	0.97	0.88	0.93	0.88	0.92	0.85
	XGboost	210	35	258	18	0.86	0.93	0.90	0.92	0.89	0.80
	LREG	213	32	264	12	0.87	0.96	0.92	0.95	0.91	0.83
HIV/AIDS	PLS-DA	15	0	1310	46	1.00	0.97	0.97	0.25	0.39	0.49
	ANN	12	3	1273	83	0.80	0.94	0.94	0.13	0.22	0.30
	KNN	12	3	1250	106	0.80	0.92	0.92	0.10	0.18	0.27
	SVM	11	4	1261	95	0.73	0.93	0.93	0.10	0.18	0.26
	SIMCA	13	2	1213	143	0.87	0.89	0.89	0.08	0.15	0.25
	XGboost	15		1255	101	1.00	0.93	0.93	0.13	0.23	0.35
	LREG	11	4	1236	120	0.73	0.91	0.91	0.08	0.15	0.23
TB	PLS-DA	31	3	1251	105	0.91	0.92	0.92	0.23	0.36	0.43
	ANN	29	5	1247	109	0.85	0.92	0.92	0.21	0.34	0.40
	KNN	29	5	1247	109	0.85	0.92	0.92	0.21	0.34	0.40
	SVM	31	3	1220	136	0.91	0.90	0.90	0.19	0.31	0.39
	SIMCA	25	9	1207	149	0.74	0.89	0.89	0.14	0.24	0.29
	XGboost	29	5	1251	105	0.85	0.92	0.92	0.22	0.35	0.41
	LREG	24	10	1257	99	0.71	0.93	0.92	0.20	0.31	0.34
	PLS-DA	22	2	1343	13	0.92	0.99	0.99	0.63	0.75	0.75
	ANN	21	3	1255	101	0.88	0.93	0.92	0.17	0.29	0.37
HIV/TB	KNN	20	4	1238	118	0.83	0.91	0.91	0.14	0.25	0.33
	SVM	18	6	1243	113	0.75	0.92	0.91	0.14	0.23	0.30
	SIMCA	11	13	1214	142	0.46	0.90	0.89	0.07	0.12	0.15
	XGboost	20	4	1252	104	0.83	0.92	0.92	0.16	0.27	0.35
	LREG	19	5	1273	83	0.79	0.94	0.94	0.19	0.30	0.36

Group of patients	Total	TP	FN	Non-error rate	Error rate
Diabetes	86	85	1	0.988	0.012
Deslipidemias^a	282	275	7	0.975	0.025
Hypothyroidism	184	181	3	0.984	0.016
Obesity	676	664	12	0.982	0.018
Total	1228	1205	23	0.981	0.019

Sociedade Brasileira de Química Instituto de Química - UNICAMP, Caixa Postal 6154, 13083-970 Campinas SP - Brazil, Tel./FAX.: +55 19 3521-3151 - São Paulo - SP - Brazil
E-mail: office@jbcs.sbq.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] * e-mail: pontarolo@ufpr.br