Table 1

Therapy Classification

Standard Datenum

Incremental Reduction

Bimodal Classification


R

NR

Mean AUC

R

NR

Mean AUC

R

NR

Mean AUC


AZT

526

390

0.7750

581

335

0.8550

395

521

0.7802


AZT, IDV

182

148

0.7803

189

141

0.9281

144

186

0.9107


DDI

466

273

0.7572

503

236

0.8363

272

467

0.7648


DDI, NFV

249

130

0.7352

264

115

0.8004

175

204

0.6814


D4T

450

307

0.7654

482

275

0.8081

274

483

0.7683


D4T, NFV

266

153

0.7499

280

139

0.6664

181

238

0.6713


D4T, NFV

372

200

0.7377

391

181

0.8455

260

312

0.7613


D4T, DDI, NFV

234

115

0.7518

242

107

0.7764

173

176

0.6817


3TC

582

466

0.7721

654

394

0.9280

408

640

0.7788


3TC, IDV

187

151

0.7748

196

142

0.9030

144

194

0.8763


3TC, NFV

202

159

0.7535

242

119

0.8810

175

186

0.8606


3TC, AZT

509

379

0.7731

560

328

0.8439

391

497

0.7845


3TC, AZT, IDV

177

145

0.7849

184

138

0.8858

144

178

0.8815


DDI, EFV

248

121

0.7389

208

89

0.9312

192

177

0.6711


D4T, EFV

260

125

0.7406

285

100

0.8479

194

191

0.9887


D4T, DDI, EFV

233

107

0.7516

254

86

0.9446

188

152

0.7499


3TC, EFV

207

130

0.7313

245

100

0.9731

179

166

0.9497


All Therapies

1115

904

0.7644

1188

831

0.8351

700

1319

0.8402


The overall statistics of the clinically annotated reverse transcriptase sequences from the Stanford HIV-1 Drug Resistance Database. The table shows breakdown of patients in each therapy regimen using the three different classification rules: Standard Datenum (SD), Incremental Reduction (IR), and Bimodal Classification (BM). R; responders, NR; non responders. The average AUC over 500 training/testing iterations indicate the success in differentiating responders from non responders using short linear sequence motifs as features in machine learning.

Dampier et al. BMC Medical Genomics 2009 2:47   doi:10.1186/1755-8794-2-47

Open Data