What purpose is NGS mostly used for?

I recently directed this question to pharmaceutical researchers. The answers are for as follows:

Drug target discovery - 28%:

In my view, this is the most feasible use of NGS for pharmaceutical research. NGS is a high-throughput, unbiased method giving you a ton of information about the genomic landscape. Whether you are comparing disease vs healthy, or treated vs nontreated conditions, it is giving you a broad picture. Most widely used information from the output of an NGS experiment is up or downregulated genes or pathways. As a technology that has been around for quite a while, with certain exceptions, the data processing pipelines are pretty streamlined. 

In this context, it is a perfect starting point for searching for a target and generating a hypothesis. It has certain drawbacks though. By nature, as it profiles transcriptional (mRNA) profiles, this information cannot be directly translated into the protein domain. In other words, a gene being differentially expressed does not directly imply that its protein product is differentially expressed and vice versa. This may be due to the nature of the biology (mRNA degradation, etc) or technical limitations. A second limitation is that NGS experiments can be less sensitive to differences for lowly expressed genes, such as some transcription factors. As a result, some differentially expressed genes can be missed. Therefore, it may be a good idea to complement NGS with other techniques, including proteomics methods.

Drug target validation - 44%:

It is a bit surprising to see that people are widely using NGS for drug target validation. In fact, this is technically feasible. You may want to see the effects of a drug candidate on its target gene and NGS may help you with that. However, to assess the effects on the direct target, a targeted experiment like qPCR can be more effective, as the sensitivity is higher. And this can be complemented by a proteomic experiment. In contrast, results derived from NGS data are often open to interpretation - it usually does not confirm or validate a drug target or hypotheses directly. 

Hence, although it is technically feasible to use NGS for validation of a drug target, in most cases, NGS experiment result does not give a yes/no answer. Instead, it gives you a picture which you can view, think and make a conclusion about it. That's why it is a less than ideal approach for drug target validation, though it is still plausible.

Study downstream targets - 22%:

For the studying drug targets (downstream targets of the direct target) NGS can be helpful. But from my personal experience, I noticed that indirect targets are difficult to capture with NGS. The reason is that, in most cases, the indirect targeting takes place through mechanisms like phosphorylation after gene transcription or translation. These downstream effects are not possible to capture with NGS, as it can only profile transcription.

One exception is when the direct target of the drug is a transcription factor (TF), which turns the transcription event on and off. In this case, targeting a TF can change the expression of its downstream targets. Even in this case, it may be challenging to observe the effect on indirect targets (downstream targets of the TF), especially for the positive regulation events, as there are often backup mechanisms in the cell, which can compensate the loss of activity of the target TF. In other words, if a transcription factor activates gene expression, inhibiting that TF may not necessarily downregulate its targets, as it can possibly be compensated for by other TFs. However, if a TF has inhibitory effects on its downstream targets, effects of inhibiting this TF can possibly be observed in NGS data.

Other - 6%:

Other than these purposes, NGS data can be used for drug repurposing, together with other tools and databases such as NIH LINCS and CMap, although the effectiveness might be limited.