Frequently Asked Questions

What is SMEpred?

SMEpred is the first online workbench to design and check the activity/efficacy of the chemically modified siRNA(s) for inhibiting the gene against which it is targeted i.e. how much efficiently your cm-siRNA will inhibit the gene after chemical modifications.

What is need of SMEpred webserver?

Experimental validation of cm-siRNAs needs plenty of resources. Additionally siRNA permutations and combination of different chemical modifications on different positions becomes very high e.g. if 25 different modifications will be tested on single siRNA with 21 positions only once (42 positions for two strands; sense-antisense) it will generate 1050 (25 x 42) different siRNAs which is a tedious job to choose from. If we combine more than one modifications on same siRNA, the permutations and combinations increases exponentially which is tough job near to impossible for any lab to perform in terms of cost, time and human resource. SMEpred predicts efficacy score computationally for the cm-siRNA sequence(s). This webserver allows the user to check the efficacy of any number of permutations and combinations for defined chemical modifications.

To whom is SMEpred useful?

SMEpred webserver is useful to all the researchers whose work involves siRNA or cm-siRNAs either for general research purposes or therapeutics development.

Can user find the list of chemical modifications used on the algorithm page ?

Yes, for the convenience we have provided list of the chemical modifications and the one letter code for the same.

What are the components of SMEpred?

SMEpred contains three components namely SMEpred, MultiModGen and siMEpred-tool besides other pages as Help-page, FAQS, Contact etc.

Does SMEpred predicts the efficacy of normal (unmodified) siRNA?

Yes, along with the cm-siRNAs efficacy SMEpred do predict the efficacy of unmmodified siRNAs. This algoithm is derived from the previous work of our lab namely siRNApred ( built to check the efficacy of siRNA(s).

Where can I find list of chemically modified siRNA?

List of chemically modified siRNA is provided as the list in the HELP PAGE. It contains the usual 5 nucleic acids Adenosine (A) Guanosine (G) Cytosine (C) Thymidine (T) Uridine Nucleic Acid (U) and 30 modified nucleic acids(click here to see the modifications). User can download the list via the link given on the same page. Futher for exploring more cm-siRNAs you can visit siRNAmod: A database of experimentally validated chemically modified siRNAs (

How was data collected for SMEpred?

We collected and curated experimentally validated data of chemically modified siRNA sequences from "siRNAmod: A database of experimentally validated chemically modified siRNAs" ( The dataset used in the model development is provided on appropriate page (Dataset).

What are the two datasets used?

We have developed SMEpred using experimentally validated chemically modified siRNAs (two datasets). These data sets is collected and curated from the literature. First dataset (Hetero-3031) have 3031 chemically modified siRNAs tested under heterogeneous experimental conditions involving up to 30 chemical modifications at 1 to 21 siRNA nucleotide positions. This dataset is divided into train/test set (Hetero-T2727) and validation set (Hetero-V304). Second dataset (Homo-2110) consists of 20 different chemical modifications experimentally verified in homogenous conditions in a single study. For Homo-2110, we used Homo-T1900 as train/test set and Homo-V210 as validation set.

How was SMEpred model built?

3031 unique chemically modified siRNA entries with 30 most commonly used chemical modifications were used (for which data was appropriate to built model). These entries along with their activity/efficiency to silence their targets were used to build the model. The features used were nucleotide frequency (mono, di), binary pattern, mono + binary pattern, mono-binary-N12C12 terminal.

What is the difference between three models of siMEpred-tool?

Differen models of this tool are based on different training features. Model-A is based on mononucleotide composition (MNC) and showed the highest PCC during model development. Model-B is the combination of MNc and binary pattern (defining the position of nucleotides adn modified nucleotides) of forst 13-nucleotides comprising of seed region of antisense strand only. Model-C consists of MNc and binary of last 8 nucleotides startin from 3'-end of anisese srand. thsi tool is provided to explore the position of modification on the antisense strand as this strand is finally utilized in RNAi machinary and silencing target gene.