The read-across and trend analysis (i.e. based of searched analogues) could be applied only on single chemical.
You can make a prediction for a list of chemicals by any of the available (Q)SAR models as well as by the automated workflows (two automated workflows - associated with skin sensitization and aquatic toxicity, are available currently).

In order to make a prediction for the isomers of interest they all should be on the data matrix. For this purpose you could: 1) load a .smi file containing the isomers or 2) to add them in a category one by one (Figure 1)
Figure 1

Once all chemicals are on the data matrix, go to the Data gap filling module. Select a row of the matrix which corresponds to the endpoint of interest (e.g. genetic toxicity) (1) and  click on the QSAR button (2) (Figure 2). All relevant QSAR models will appear. Select a model (3) and click on Run (4). Choose to predict All chemicals (5) and confirm by Ok (6).

Figure 2