| 查看: 262 | 回复: 0 | ||
huqionghhkk木虫 (小有名气)
|
[求助]
急急急！！！谢谢，求一段翻译
|
|
For each compound we constructed a performance profile assigning binary ([0,1], binding data) or discrete (½−1;0;1 , functional data) values representing activity for a compound in a given assay, collecting such values across all assays into a vector x. Discretization of binding data are described elsewhere (48), and ChemBank data were handled similarly, except both high- and low-signal outlier values (56) were accepted (SI Datasets D6 and D7). All distinct performance profile vectors x for compounds were collected to set S, and Shannon entropy (H) was computed by calculating relative frequencies pðxÞ and summing frequency terms over x ∈ S: HðSÞ ¼ −ΣpðxÞ log2½pðxÞ (profile entropy). To calculate entropy for a set of profiles weighted by a specificity constraint, we first computed H separately for each subset S m of profiles sharing the same number m of nonzero profile features, then computed a weighted sum of these entropies with weighting factor w m ¼ exp½− lnð2Þm , to give H w ðSÞ ¼ Σw m HðS m Þ (weighted profile entropy). For ChemBank profiles, missing data were censored to zero before entropy calculations (SI Dataset D8). Statistical analyses, visualizations, and entropy calculations were performed in MATLAB. |
回复此楼» 猜你喜欢
各位大神，目前国内有哪些比较好用的逆合成软件？
已经有14人回复
-
国社科系统bug了，是不是要放榜了？
已经有10人回复
-
现代”学阀”该如何界定
已经有9人回复
-
课题组招2027级博士 上海工程技术大学 激光智能制造方向
已经有4人回复
-
上海工程技术大学 激光智能制造课题组 2027级博士研究生招生
已经有3人回复
-
上海工程技术大学激光智能制造课题组｜2027级博士研究生招生公告
已经有11人回复
-
申博发邮件
已经有11人回复
-
我的奶奶
已经有3人回复