Table 1 Training set isolate phenotypes.

From: A convolutional neural network highlights mutations relevant to antimicrobial resistance in Mycobacterium tuberculosis

Drug

Resistant (n)

Susceptible (n)

Total (n)

Resistant proportion

Isoniazid

4232

5723

9955

0.425

Rifampicin

3472

6428

9900

0.351

Ethambutol

2273

6390

8663

0.262

Pyrazinamide

1505

5393

6898

0.218

Streptomycin

2643

4362

7005

0.377

Amikacin

773

2632

3405

0.227

Capreomycin

737

2838

3575

0.206

Kanamycin

796

2502

3298

0.241

Ciprofloxacin

118

388

506

0.233

Ofloxacin

912

2246

3158

0.289

Moxifloxacin

398

1941

2339

0.170

Levofloxacin

66

189

255

0.259

Ethionamide

791

1647

2438

0.324

Total isolates

  

10,201

 
  1. Phenotypic summary of the 10,201 isolates used to train and cross-validate the models: the numbers of resistant isolates, susceptible isolates, the total tested (sum of the numbers of resistant and susceptible isolates), and the resistant proportion, with respect to each of the 13 anti-TB drugs.
Back to article page