CovidPrognosis: A webserver for Clinical Information based COVID-19 Prognosis Prediction

Feature selection techniques applied for feature selection from the day 0, 3 and 7 clinical data.

Sr.No.	Attribute Evaluator	Search Method	Selected Features
1	CfsSubsetEval	BestFirst	age_cat, abs_lymph_0, creat_0
2	CfsSubsetEval	GreedyStepwise	age_cat, abs_lymph_0, creat_0
3	ClassifierAttributeEval	Ranker	ldh_7, CRP_0, fever_symp, GI_symp, abs_neut_0, abs_lymph_0, abs_mono_0, resp_symp, immuno, hypertension (Top-10 ranked attributes)
4	CorrelationAttributeEval	Ranker	age_cat, abs_lymph_0, heart, kidney, creat_0, hypertension, ddimer_0, trop_72, GI_symp, fever_symp (Top-10 ranked attributes)
5	GainRatioAttributeEval	Ranker	age_cat, abs_lymph_0, creat_0, heart, kidney, trop_72, hypertension, ddimer_0, GI_symp, abs_neut_0 (Top-10 ranked attributes)
6	InfoGainAttributeEval	Ranker	age_cat, abs_lymph_0, creat_0, heart, hypertension, kidney, ddimer_0, GI_symp, trop_72, immuno (Top-10 ranked attributes)
7	OneRAttributeEval	Ranker	All the attributes almost equally ranked, hence, no advantage of using this.
8	ReliefFAttributeEval	Ranker	hypertension, fever_symp, age_cat, GI_symp, diabetes, heart, creat_0, ldh_3, CRP_3, CRP_7 (Top-10 ranked attributes)
9	SymmetricalUncertAttributeEval	Ranker	age_cat, abs_lymph_0, creat_0, heart, kidney, hypertension, trop_72, ddimer_0, GI_symp, abs_neut_0 (Top-10 ranked attributes)
10	PrincipalComponents	Ranker	Not Applicable

Feature selection techniques applied for feature selection from the day 0 proteomics data.

Sr.No.	Attribute Evaluator	Search Method	Selected Features/Proteins	Features common with CfsSubsetEval and BestFirst	Number of features in common
1	CfsSubsetEval	BestFirst	P01127, P01135, P01591, P10145, P12034, P26022, P58294, Q8TD46, Q96PL1, Q9P0M4, Q9Y3P8, O14793, O60635, O95445, O95841, P00533, P02462, P04070, P08709, P09496, P14555, P36222, P55259, Q13231, Q14515, Q14767, Q16769, Q76LX8, Q92820, Q9UGM5, O76076, P13385, P13500, P13611, P22676, P55291, Q15633, Q8IU54, Q96NZ8, O15123, P29017, P35968, P50749, Q8NBZ7, Q9Y653	Not Applicable	Not Applicable
2	CfsSubsetEval	GreedyStepwise	P01127, P01135, P01591, P10145, P12034, P26022, P58294, Q8TD46, Q96PL1, Q9P0M4, Q9Y3P8, O14793, O60635, O95445, O95841, P00533, P02462, P04070, P08709, P09496, P14555, P36222, P55259, Q13231, Q14515, Q14767, Q16769, Q76LX8, Q92820, Q9UGM5, O76076, P13385, P13500, P13611, P22676, P55291, Q15633, Q8IU54, Q96NZ8, O15123, P29017, P35968, P50749, Q8NBZ7, Q9Y653	Q96NZ8, O15123, P29017, P35968, P50749, Q8NBZ7, P26022, Q9Y653, P58294, Q8TD46, Q96PL1, Q9P0M4, Q9Y3P8, O14793, O60635, O95445, O95841, P00533, P02462, P04070, P08709, P09496, P14555, P36222, P01127, P55259, P01135, Q13231, Q14515, Q14767, Q16769, Q76LX8, P01591, Q92820, Q9UGM5, O76076, P13385, P13500, P13611, P22676, P10145, P12034, P55291, Q15633, Q8IU54	45
3	ClassifierAttributeEval	Ranker	Q9Y6A5, P13501, P13591, P12724, P12830, P12318, P16109, P10721, P12104, P12111, P13598, P13686, P13807, P15144, P15529, P13987, P15090, P15086, P15085, P14555, P14543, P10646, P10644, P10586, P08833, P08887, P09237, P08709, P08670, P08581, P08571, P08319, P09093, P09382, P10451, P09668, P10145, P09417, P09619, P09601, P09525, P09496, P09467, P15907, P16112, P31483, P22748, P23141, P21980, P22004 (Top-50 ranked attributes)	P08709, P09496, P14555	3
4	CorrelationAttributeEval	Ranker	Q9P0M4, P52823, Q14767, P07585, P51888, O14558, P35318, P16234, P13611, P36222, P20827, P08833, Q9Y653, Q9BXY4, Q2MKA7, Q9BU40, Q9UGM5, O76076, O95633, Q9Y279, Q9HAV5, Q9HCB6, Q8TCT1, Q6UX15, Q14508, Q9UBP4, Q9NQ30, P15090, P19438, Q9Y5K2, Q96PL1, P29475, P62736, P78423, P10451, P02144, P29317, O43570, O15232, P01135, Q9UHD0, Q9UM47, O00300, Q9NP84, Q9NZV1, P07196, P24821, Q8NI22, P55259, Q5ZPR3 (Top-50 ranked attributes)	Q9Y653, Q96PL1, Q9P0M4, P36222, P55259, P01135, Q14767, Q9UGM5, O76076, P13611	10
5	GainRatioAttributeEval	Ranker	Q76LX8, P08709, Q96PL1, P13611, Q8NBZ7, P09238, P35318, Q9BXY4, Q9HAV5, P07196, Q14767, P01135, Q8IX05, O00300, P07585, Q9Y6Q6, P51888, P36222, Q9P0M4, P16112, Q9Y5W5, Q14508, O15232, P62736, O14558, P01591, Q01973, P22223, Q9Y653, P26022, Q9H3U7, P98160, Q2MKA7, O14763, P15090, Q13275, P08833, Q8IU54, O14793, P20827, Q6UX15, Q5ZPR3, P04080, P49763, Q8TCT1, Q8WXI8, Q8N9I9, P24821, Q9NS68, P43121 (Top-50 ranked attributes)	Q8NBZ7, P26022, Q9Y653, Q96PL1, Q9P0M4, O14793, P08709, P36222, P01135, Q14767, Q76LX8, P01591, P13611, Q8IU54	14
6	InfoGainAttributeEval	Ranker	Q9Y653, P52823, P36222, Q14767, Q9UGM5, P13611, Q8TCT1, O14558, Q96PL1, Q9P0M4, P08833, P51888, P49763, O76076, Q2MKA7, Q6UX15, P04080, P20827, P07585, P55259, P19438, Q14508, O95633, Q9GZV9, O00300, Q9NP84, Q8NI22, Q96FE7, Q9HAV5, Q9BXY4, P62736, Q9NQX5, P26022, P35318, P15090, Q08708, Q9HCB6, P10451, P02144, Q9H3U7, P98160, Q9Y5K2, P13987, P29317, P19957, P19429, Q9NZV1, O14904, P24821, Q9NS68 (Top-50 ranked attributes)	P26022, Q9Y653, Q96PL1, Q9P0M4, P36222, P55259, Q14767, Q9UGM5, O76076, P13611	10
7	OneRAttributeEval	Ranker	Q9P0M4, Q8NBZ7, Q9BU40, P14868, O95715, Q01973, Q6NW40, P12724, P24821, Q9BZM5, Q8IX05, Q8TDQ1, Q9UGM5, Q9HAV5, O00622, Q99650, Q9HCB6, Q9NQX5, P35237, Q13219, P13611, P13500, P07204, P09238, Q96B86, O75144, Q96LC7, P14778, Q8WTT0, P15018, P01135, P29317, Q15517, Q9UP79, Q9Y5W5, P21246, P15328, O95388, Q9BSG5, P12104, Q6UX15, P21709, Q8NBS9, Q14126, P01375, Q12933, O94903, P04216, Q96J42, Q9H6B4 (Top-50 ranked attributes)	Q8NBZ7, Q9P0M4, P01135, Q9UGM5, P13500, P13611	6
8	ReliefFAttributeEval	Ranker	Q9Y653, P36222, O95866, P35968, P08727, Q6PJW8, O15117, Q8NDB2, P35318, O43561, P08134, Q9Y6K9, P42574, O14974, Q9HD26, O75563, Q9P1Z2, Q07108, O00161, P01133, Q96SB3, Q9UM47, Q8NEZ2, O60884, Q9NZN5, P01135, P10644, Q9UII2, Q9BXY4, P07585, P07948, P10606, P29965, Q8IU54, P16860, Q14767, O14558, Q99685, P07947, P47712, P11274, O95786, Q9Y4K4, Q8TCT1, Q9GZM7, O43597, Q9NQ30, O95721, P04070, Q9UGM5 (Top-50 ranked attributes)	P35968, Q9Y653, P04070, P36222, P01135, Q14767, Q9UGM5, Q8IU54	8
9	SymmetricalUncertAttributeEval	Ranker	Q96PL1, P13611, Q14767, P36222, Q9P0M4, Q9Y653, P35318, P51888, Q9BXY4, Q9HAV5, O14558, P07585, O00300, Q14508, P01135, Q8IX05, Q2MKA7, P62736, P08833, Q8TCT1, P07196, P09238, Q9Y6Q6, P26022, Q6UX15, P20827, P49763, P04080, Q9H3U7, P98160, P15090, O15232, O76076, Q9GZV9, Q9UGM5, Q96FE7, P52823, P02144, Q9HCB6, P10451, Q9NS68, P24821, Q9NQX5, Q13275, Q8N9I9, Q8WXI8, P12111, Q8NBZ7, O14763, P19957 (Top-50 ranked attributes)	Q8NBZ7, P26022, Q9Y653, Q96PL1, Q9P0M4, P36222, P01135, Q14767, Q9UGM5, O76076, P13611	11
10	PrincipalComponents	Ranker	Not Applicable	Not Applicable	Not Applicable

Result files containing models evaluation performance for Clinical parameters and Proteomics information-based COVID-19 Prognosis prediction models.

Clinical information based models: Day 0 (all clinical parameters based)
1.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using whole dataset.	clinical_info_only_day_0_all_params.csv
2.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P1.	d_0_all_params_x00.csv
3.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P2.	d_0_all_params_x01.csv
4.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P3.	d_0_all_params_x02.csv
5.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P4.	d_0_all_params_x03.csv
6.	Results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P5.	d_0_all_params_x04.csv
7.	Average of the results for clinical information based models (trained and evaulated with day 0 all clinical parameters) developed using split P1-P5.	avg_res_clinical_info_only_day_0_all_params.csv
Clinical information based models: Day 3 (all clinical parameters based)
8.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using whole dataset.	clinical_info_only_day_3_all_params.csv
9.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P1.	d_3_all_params_x00.csv
10.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P2.	d_3_all_params_x01.csv
11.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P3.	d_3_all_params_x02.csv
12.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P4.	d_3_all_params_x03.csv
13.	Results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P5.	d_3_all_params_x04.csv
14.	Average of the results for clinical information based models (trained and evaulated with day 3 all clinical parameters) developed using split P1-P5.	avg_res_clinical_info_only_day_3_all_params.csv
Clinical information based models: Day 7 (all clinical parameters based)
15.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using whole dataset.	clinical_info_only_day_7_all_params.csv
16.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P1.	d_7_all_params_x00.csv
17.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P2.	d_7_all_params_x01.csv
18.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P3.	d_7_all_params_x02.csv
19.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P4.	d_7_all_params_x03.csv
20.	Results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P5.	d_7_all_params_x04.csv
21.	Average of the results for clinical information based models (trained and evaulated with day 7 all clinical parameters) developed using split P1-P5.	avg_res_clinical_info_only_day_7_all_params.csv
Clinical information based models: Day 0,3,7 (all clinical parameters based)
22.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using whole dataset.	clinical_info_only_day_0_7_all_params.csv
23.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P1.	d_0_7_all_params_x00.csv
24.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P2.	d_0_7_all_params_x01.csv
25.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P3.	d_0_7_all_params_x02.csv
26.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P4.	d_0_7_all_params_x03.csv
27.	Results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P5.	d_0_7_all_params_x04.csv
28.	Average of the results for clinical information based models (trained and evaulated with day 0,3,7 all clinical parameters) developed using split P1-P5.	avg_res_clinical_info_only_day_0_7_all_params.csv
Clinical information based models: Day 0,3,7 (selected 3 clinical parameters based)
29.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using whole dataset.	clinical_info_only_day_0_7_sel_3_features.csv
30.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P1.	d_0_7_sel_3_params_x00.csv
31.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P2.	d_0_7_sel_3_params_x01.csv
32.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P3.	d_0_7_sel_3_params_x02.csv
33.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P4.	d_0_7_sel_3_params_x03.csv
34.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P5.	d_0_7_sel_3_params_x04.csv
35.	Average of the results for clinical information based models (trained and evaulated with day 0,3,7 selected 3 clinical parameters; BestFirst algorithm-based) developed using split P1-P5.	avg_res_clinical_info_only_day_0_7_sel_3_features.csv
Clinical information based models: Day 0,3,7 (selected 9 clinical parameters based)
36.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using whole dataset.	clinical_info_only_day_0_7_sel_9_features.csv
37.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P1.	d_0_7_sel_9_params_x00.csv
38.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P2.	d_0_7_sel_9_params_x01.csv
39.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P3.	d_0_7_sel_9_params_x02.csv
40.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P4.	d_0_7_sel_9_params_x03.csv
41.	Results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P5.	d_0_7_sel_9_params_x04.csv
42.	Average of the results for clinical information based models (trained and evaulated with day 0,3,7 selected 9 clinical parameters; Ranker algorithm-based) developed using split P1-P5.	avg_res_clinical_info_only_day_0_7_sel_9_features.csv
Proteomics information based models: Day 0 (using all 1428 proteins NPX values)
43.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using whole dataset.	covid_D0_NPX_all_proteins.csv
44.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P1.	covid_D0_NPX_all_proteins_x00.csv
45.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P2.	covid_D0_NPX_all_proteins_x01.csv
46.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P3.	covid_D0_NPX_all_proteins_x02.csv
47.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P4.	covid_D0_NPX_all_proteins_x03.csv
48.	Results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P5.	covid_D0_NPX_all_proteins_x04.csv
49.	Average of the results for proteomics information based models (trained and evaluated with day 0 all 1428 proteins NPX values) using split P1-P5.	avg_res_covid_D0_NPX_all_proteins.csv
Proteomics information based models: Day 0 (using selected 45 proteins NPX values)
50.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using whole dataset.	Day_0_NPX_sel_45_proteins_all_records.csv
51.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P1.	Day_0_sel_45_NPX_x00.csv
52.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P2.	Day_0_sel_45_NPX_x01.csv
53.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P3.	Day_0_sel_45_NPX_x02.csv
54.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P4.	Day_0_sel_45_NPX_x03.csv
55.	Results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P5.	Day_0_sel_45_NPX_x04.csv
56.	Average of the results for proteomics information based models (trained and evaluated with day 0 selected 45 proteins NPX values) using split P1-P5.	avg_res_Day_0_NPX_sel_45_proteins_all_records.csv