Development of an ecotoxicity QSAR model for the KAshinhou Tool for Ecotoxicity (KATE) system, March 2009 version

Figures & data

Table 1. QSARs for fish acute toxicity estimated by the equation: log(1/LC ₅₀[mM]) = a * log P + b.

Class name	a, b	n	RMSE	r^2, q^2	log P range	^*1
Hydrocarbons aromatic	0.630, −0.883	43	0.368	0.826, 0.803	[0.60, 5.17]
Dinitrobenzenes	0.568, 0.551	12	0.669	0.331, 0.170	[0.56, 3.60]
Nitrobenzenes	0.678, −0.693	9	0.300	0.875, 0.760	[0.82, 5.10]
Amines aromatic or phenols1	−0.005, 2.671	7	0.354	0.001, 0.887	[−0.30, 4.47]
Amines aromatic or phenols2	0.012, 1.863	7	0.307	0.003, 0.737	[3.67, 8.47]	C
Amines aromatic or phenols3	0.214, 0.945	16	0.305	0.272, 0.106	[0.15, 3.68]
Amines aromatic or phenols4	0.725, −0.779	56	0.321	0.900, 0.890	[0.51, 7.54]
Amines aromatic or phenols5	0.544, −0.612	22	0.324	0.661, 0.600	[0.35, 3.50]
Primary amines	0.529, −0.622	23	0.406	0.803, 0.741	[−2.04, 3.60]
Secondary and tertiary amines	0.592, −0.595	10	0.512	0.731, 0.605	[−1.43, 2.79]	C
Hydrazines	0.417, 1.832	4	0.413	0.884, 0.639	[−1.68, 4.70]
Amides and imides	0.746, −1.026	17	0.601	0.696, 0.607	[−0.48, 3.80]
Esters aliphatic	0.638, −0.600	13	0.393	0.722, 0.651	[0.18, 3.65]
Esters aromatic	0.513, −0.157	9	0.253	0.856, 0.790	[1.94, 5.53]
Aldehydes	0.484, 0.279	15	0.557	0.272, 0.111	[−0.34, 2.47]
Acids	0.728, −1.652	9	0.355	0.816, 0.667	[0.33, 4.20]
Acids acrylic	0.122, 0.045	3	0.039	0.607, 0.271	[0.35, 1.33]
Conjugated systems1	0.753, 2.084	4	1.012	0.463, 0.111	[−1.11, 2.20]
Conjugated systems2	0.436, 0.901	17	1.007	0.264, 0.066	[−0.38, 4.10]
Thiols aromatic	NO-QSAR
Thiols aliphatic	0.371, 0.732	4	0.291	0.910, 0.633	[−0.17, 6.12]	C
Sulfides	0.753, −1.336	8	0.259	0.699, 0.573	[2.46, 4.16]	C
Disulfides	0.386, 0.845	6	0.666	0.210, 0.012	[1.74, 4.44]	C
Carbamates	0.004, 1.894	11	0.519	0.000, 0.645	[−0.47, 4.60]
Pyrethroids	NO-QSAR
Acrylates	0.158, 1.498	6	0.155	0.474, 0.022	[−0.21, 2.36]
Methacrylates	0.465, −0.031	6	0.417	0.657, 0.293	[0.47, 4.54]
Epoxides	0.323, 1.055	4	0.272	0.755, 0.283	[0.08, 3.98]	C
Barbitals or thiols other	1.583, −2.560	4	0.291	0.657, 0.927	[1.47, 2.10]
Esters phosphate	0.691, −0.111	11	0.856	0.389, 0.175	[2.23, 5.33]
N or P cations	0.274, 0.956	9	0.579	0.791, 0.628	[−8.36, 6.69]	C
Halides1	0.254, 1.325	6	0.971	0.112, 0.078	[0.45, 4.50]
Halides2	0.824, −0.318	8	0.560	0.879, 0.810	[−0.06, 5.04]
Halides3	0.783, −1.291	42	0.263	0.879, 0.868	[1.25, 4.89]
Metals	NO-QSAR
Nitriles aliphatic	0.839, −1.154	6	0.254	0.938, 0.901	[−0.34, 3.12]	N
Ketones	0.864, −1.602	21	0.345	0.891, 0.867	[−0.24, 4.09]	N
Alcohols or ethers aliphatic	0.853, −1.958	23	0.321	0.950, 0.924	[−0.77, 5.82]	N
Phosphates	0.891, −1.926	3	0.257	0.865, 0.485	[2.83, 4.59]	N
Hydrocarbons aliphatic	0.753, −1.286	15	0.289	0.824, 0.785	[2.42, 5.56]	N
Ethers aliphatic	0.749, −1.806	11	0.190	0.972, 0.962	[−0.54, 4.25]	N
Ethers aromatic	0.870, −1.466	10	0.233	0.922, 0.892	[1.16, 4.21]	N
Neutral organics	0.842, −1.674	88	0.384	0.924, 0.919	[−0.77, 5.82]
Unclassified	0.744, −0.898	25	0.714	0.712, 0.660	[−1.35, 5.50]
^*1 C: an equation is generated by calculated Clog P. N: a member of the Neutral organics class.
Note: n, RMSE, r ^2 and q ^2 denote the number of chemicals in a class, the root mean square error, the squared correlation coefficient, and the leave-one-out version of the squared correlation coefficient, respectively. The log P range shows minimum and maximum log P values.

Table 2. QSARs for the daphnia acute toxicity estimated by the equation: log(1/EC ₅₀[mM]) = a * log P + b.

Class name	a, b	n	RMSE	r ^2, q ^2	log P range	^*1
Hydrocarbons aromatic	0.607, −0.414	26	0.351	0.808, 0.762	[0.65, 5.17]
Dinitrobenzenes	0.408, 0.632	5	0.561	0.343, 0.090	[0.56, 3.60]
Nitrobenzenes	0.547, −0.164	4	0.238	0.915, 0.675	[1.17, 5.10]
Amines aromatic or phenols1	0.085, 2.441	7	0.443	0.057, 0.375	[−0.33, 3.41]
Amines aromatic or phenols2	0.097, 1.152	6	0.277	0.239, 0.031	[3.67, 8.47]	C
Amines aromatic or phenols3	0.132, 1.748	16	0.406	0.119, 0.018	[0.04, 3.91]
Amines aromatic or phenols4	0.576, −0.042	28	0.297	0.838, 0.814	[1.32, 6.06]
Amines aromatic or phenols5	0.552, 0.114	12	0.260	0.802, 0.728	[1.18, 3.91]
Primary amines	0.189, −0.059	4	0.248	0.390, 0.095	[−1.31, 1.49]
Secondary and tertiary amines	0.133, 0.200	4	0.150	0.517, 0.040	[−1.50, 1.45]
Hydrazines	0.190, 1.987	5	0.289	0.766, 0.360	[−2.46, 4.70]	C
Amides and imides	0.212, 0.585	8	0.593	0.151, 0.135	[0.23, 3.80]
Esters aliphatic	0.666, −0.819	6	0.324	0.927, 0.762	[0.25, 5.41]
Esters aromatic	0.459, −0.417	3	0.010	1.000, 0.998	[1.60, 4.72]
Aldehydes	0.521, 0.295	5	0.555	0.616, 0.084	[0.42, 4.47]	C
Acids	0.222, −0.113	7	0.644	0.133, 0.298	[0.08, 4.20]
Acids acrylic	0.057, 0.248	3	0.143	0.025, 0.947	[0.35, 1.33]
Conjugated systems1	0.630, 1.393	5	0.321	0.957, 0.916	[−1.76, 4.65]	C
Conjugated systems2	0.213, 0.906	11	0.775	0.097, 0.047	[0.17, 3.70]
Thiols aromatic	NO–QSAR
Thiols aliphatic	0.427, 1.410	4	0.786	0.647, 0.049	[−0.17, 6.12]	C
Sulfides	NO–QSAR
Disulfides	1.041, −0.724	3	0.480	0.865, 0.635	[1.74, 4.44]	C
Carbamates	0.046, 2.991	4	0.688	0.008, 0.523	[0.94, 4.60]
Pyrethroids	NO–QSAR
Acrylates	0.003, 1.401	4	0.069	0.002, 0.646	[−0.21, 2.36]
Methacrylates	0.461, −0.422	5	0.301	0.824, 0.653	[0.47, 4.54]
Epoxides	0.486, 0.589	4	0.341	0.817, 0.598	[0.08, 3.98]	C
Barbitals or thiols other	NO–QSAR
Esters phosphate	2.133, −2.376	3	1.477	0.204, 0.526	[3.08, 3.88]
N or P cations	NO–QSAR
Halides1	−0.665, 4.825	3	0.350	0.800, 0.998	[2.09, 4.50]
Halides2	0.880, −0.317	4	0.552	0.860, 0.494	[1.10, 5.04]
Halides3	0.826, −1.008	24	0.237	0.901, 0.883	[1.47, 4.73]
Metals	NO–QSAR
Nitriles aliphatic	NO–QSAR					N
Ketones	NO–QSAR					N
Alcohols or ethers aliphatic	0.641, −1.053	6	0.214	0.958, 0.923	[1.10, 5.82]	N
Phosphates	0.579, −0.634	3	0.103	0.983, 0.922	[1.44, 4.59]	N
Hydrocarbons aliphatic	0.660, −0.555	10	0.268	0.891, 0.797	[2.42, 6.54]	N
Ethers aliphatic	NO–QSAR					N
Ethers aromatic	0.492, 0.285	4	0.437	0.406, 0.088	[2.16, 4.21]	N
Neutral organics	0.696, −0.870	26	0.418	0.857, 0.835	[0.68, 6.54]
Unclassified	0.537, 0.078	12	1.097	0.475, 0.287	[−1.02, 5.50]
^*1 C: an equation is generated by the calculated Clog P. N: a member of the Neutral organics class.
Note. n, RMSE, r ^2, and q ^2 denote the number of chemicals in a class, the root mean square error, the squared correlation coefficient, and the leave-one-out version of the squared correlation coefficient, respectively. The log P range shows minimum and maximum log P values.

Figure 1. The correlation between log P and the measured toxicity values of chemicals used in KATE as a daphnia end-point. The dotted-dashed, dashed and bold lines are the QSAR equations of amines aromatic or phenols4, amines aromatic or phenols5, and neutral organics, respectively.

Table 3. Statistical information comparing measured and calculated fish log(1/LC ₅₀[mM]) of 287 test set compounds. The complete results are shown in Appendix 5-1.

	TIMES^*1	ECOSAR^*2	KATE^*2
			All^*3	log P ^*4	C(1)^*5	C(2)^*6	log P ^*4 C(1)^*5	log P ^*4 C(2)^*6
Chemicals ^*7	274	242	274	207	152	192	111	144
Predicted ^*8	274	259	318	252	187	233	145	179
r ^2	0.751	0.790	0.868	0.833	0.901	0.890	0.886	0.866
RMSE	0.935	0.869	0.685	0.641	0.644	0.655	0.588	0.617
Under^*9 [%]	11.3	10.0	4.7	5.2	5.3	5.6	2.8	3.9
Over^*10 [%]	5.1	8.1	7.2	6.7	8.0	6.9	8.3	7.3
Notes:
^*1Each chemical is identified by one QSAR class.
^*2When a chemical is found to belong to more than one QSAR class, all the estimated data are adopted. If only the name of the class is available, such data are omitted.
^*3Both in-domain and out-of-domain data for log P and C-judgements are included.
^*4In-domain of log P-judgement.
^*5In-domain of C-judgement is defined as all substructures of a test chemical being found in reference chemicals in the class.
^*6In-domain of C-judgement defined as all substructures of a test chemical being in reference chemicals in either Neutral organics or the class.
^*7The number of compounds that can be predicted.
^*8The total number of the predicted values by using the training sets. Some chemicals belong to more than one class, and thus Predicted is larger than Chemicals. r ^2, RMSE, Under and Over were calculated based on the Predicted number.
^*9Fractions (%) of the underestimated chemicals. Underestimation is defined as [calculated log(1/LC 50) – measured log(1/LC 50)] < −1.
^*10Fractions (%) of the overestimated chemicals. Overestimation is defined as [calculated log(1/LC 50) – measured log(1/LC 50)] > 1.

Table 4. Statistical information between measured and calculated Daphnia log(1/EC ₅₀[mM]) for 98 test set compounds. The complete results are shown in Appendix 5-2.

	TIMES^*1	ECOSAR^*2	KATE^*2
			all^*3	log P^*4	C(1)^*5	C(2)^*6	log P ^*4C(1)^*5	log P ^*4C(2)^*6
Chemicals ^*7	93	82	94	58	43	55	25	33
Predicted ^*8	93	85	102	66	46	61	31	39
r ^2	0.668	0.699	0.662	0.732	0.793	0.686	0.807	0.801
RMSE	1.404	1.364	0.993	0.784	0.799	0.968	0.639	0.689
Under^*9 [%]	21.5	14.1	9.8	1.5	6.5	8.2	0.0	0.0
Over^*10 [%]	11.8	18.8	14.7	15.2	6.5	11.5	6.5	10.3
Notes: As in Table 3.