Table 2. TPs Which Had Another Structurally More Similar Calibration Compound than Their Parent with the RFs of TP, Parent and Most Similar Compounda.
|
parent–TP approach |
structural similarity approach |
|||||
---|---|---|---|---|---|---|---|
suspect | RFb | parent | RFb | mean fold error | most similar | RFb | mean fold error |
2-aminobenzothiazole | 5.2 × 1014 | TCMTB | 2.7 × 1013 | 29× | benzotriazole | 2.1 × 1014 | 9.1× |
benzothiazole | 3.1 × 1013 | TCMTB | 2.7 × 1013 | 44× | benzotriazole | 2.1 × 1014 | 30× |
atrazine-desethyl-desisopropyl | 4.6 × 1013 | atrazine | 1.1 × 1015 | 1 800× | simazine | 7.7 × 1014 | 1 500× |
atrazine-desethyl-desisopropyl-2-hydroxy | 3.9 × 1013 | atrazine | 1.1 × 1015 | 110× | simazine | 7.7 × 1014 | 70× |
atrazine-desisopropyl | 6.2 × 1013 | atrazine | 1.1 × 1015 | 160× | simazine | 7.7 × 1014 | 25× |
atrazine-desisopropyl-2-hydroxy | 5.6 × 1014 | atrazine | 1.1 × 1015 | 7.3× | simazine | 7.7 × 1014 | 6.0× |
metforminc | 1.0 × 1015 | guanylurea | 7.9 × 1013 | 18× | caffeine | 7.4 × 1013 | 15× |
The mean fold error, averaged over all datasets and samples for these two approaches are also shown.
RFs calculated from analysis described in SI 2 - DS_QDF, and might differ between data sets.
Metformin is actually the parent compound to guanylurea, but here they were switched due to initial model development requirements.