Differences

This shows you the differences between two versions of the page.

--- subsets [2016/12/22 21:56] – rkiss
+++ subsets [2016/12/27 21:19] – rkiss
@@ Line 1: / Line 1: @@
 ====== Prefiltered subsets of the Mcule database ======
-We provide you with Ro5 and Ro3 subsets that can serve as a starting point of your virtual screening projects if you don't want to screen the full Mcule database (36M compounds currently). Structurally diverse subsets of the drug-like and fragment-like parts were generated to represent the same chemical space with a smaller number of compounds.
+In case you cannot search / screen the full Mcule database, you may consider using some smaller, representative subsets thereof prefiltered by physicochemical properties and diversity. Structurally diverse subsets representing the drug-like (rule-of-5) and fragment (rule-of-3) chemical space can be accessed as described below.
 ===== Availability =====
 The subsets can be
-  * freely downloaded in SMILES and SDF file formats on our [[https://mcule.com/database/|download page]]
+  * freely downloaded in SMILES and SDF file formats from our [[https://mcule.com/database/|download page]]
-  * or can be selected as the input collection for [[screen|online screening]] at mcule.com if you have an [[https://mcule.com/accounts/signup/|Mcule account]]
+  * or can be selected as the input collection for [[screen|online screening]] in Mcule if you have an [[https://mcule.com/accounts/signup/|Mcule account]]
 ===== Methods =====
@@ Line 13: / Line 13: @@
 ==== Property based filtering ====
-For the drug-like and fragment-like subsets the rule of 5 and rule of 3 rules were applied, allowing one violation. Additionally, we used a few more rules:
+For the drug-like and fragment subsets the [[http://www.sciencedirect.com/science/article/pii/S0169409X00001290|rule-of-5]] and [[http://www.sciencedirect.com/science/article/pii/S1359644603028319|rule-of-3]] physicochemical property filters are applied allowing max 1 violation. Additionally, we applied the following filtering criteria to skip some rather "strange" compounds:
   * number of components < = 1
   * MW > = 100
@@ Line 20: / Line 21: @@
   * number of halogens < = 7
   * number of inorganic atoms = 0
-We used these rules to leave out more "strange compounds" which may be included more likely in the diverse sets as they differ from the regular compounds.
 ==== Diversity selection ====
-The Mcule database contains ~5.7M stock compounds and ~30.3M virtual compounds. Diversity selection was carried out in a way to prefer the stock compounds over the virtual ones. As a result, the chemical space is represented by stock compounds where possible. More over, the downloadable files contains the compounds in diversity order i.e. the first N compounds represent the most dissimilar N compounds, so you can choose yourself the number of compounds to represent the whole Ro3 and Ro5 subsets.
+Diversity selection was set up to prefer in-stock compounds over virtual ones. As a result, the chemical space is represented by in-stock compounds where possible.
+The [[https://mcule.com/database/|downloadable files]] contain the compounds in diversity order i.e. the first N compounds represent the most dissimilar N compounds. This means that if you want to further narrow down the number of compounds you can keep the first X compounds of the files and they will be the most dissimilar ones.
-Structural similarity was measured by Tanimoto coefficient (TC) between FP2 linear fingerprints generated by OpenBabel. The combinations of the following algorithms were applied to extract the most dissimilar subsets:
+Structural similarity was measured by Tanimoto coefficient (TC) between FP2 linear fingerprints generated by [[http://jcheminf.springeropen.com/articles/10.1186/1758-2946-3-33|OpenBabel]]. The combinations of the following algorithms were applied to extract the most dissimilar compounds:
-  * we used sphere exclusion to eliminate highly similar compounds to reduce the input size where needed
+  * sphere exclusion: to quickly eliminate highly similar compounds to reduce the input collection to a manageable size for the subsequent [[diversitysel|stepwise elimination]] algorithm
-  * then [[diversitysel|stepwise elimination]] was applied to obtain the most dissimilar compounds
+  * [[diversitysel|stepwise elimination]]: a more thorough algorithm that eliminates one molecule of the most similar molecule pairs
-In sphere exclusion we used the stock compounds first as "centers" for the elimination of redundant compounds, and we retained the stock compounds during the stepwise elimination.
+In sphere exclusion we used the in-stock compounds first as "centers" and eliminated their most similar analogs, while during [[diversitysel|stepwise elimination]] we retained the in-stock compounds from the most similar molecule pairs.
 ===== Subsets =====