It's not necessarily non-sense, it could work, it's just that IF is more work for the companies that prefer to focus on WB. But, just a hint: if the company has tested the antibody in some other application that requires a native protein and it worked, you have better chances of success. In particular, if the Ab works in IP, is a good sign.

Good luck

I agree with Pietro!!!

Pietro has made all the points clear. However, if the antibody is not against your molecule of interest I do not see the point in buying it.

You can always try out an antibody for use in multiple application purposes. It's just best to run a trial experiment using various concentrations of antibody in order to see if something will work. Additionally, it never hurts to call the company directly and ask if they have tested a product out for various application usage. Of course, there will be variation between the company's methods and the methods used by your laboratory.

You can test the antibody with western blotting. Also you can study a studied species with high relativity, blast the protein against the one that works with the antibody and if it shows high similarity you can assume that it could work.

What do you mean with IF? cell culture cells fixed in methanol or acetone? Actually it does not matter. Antibodies are tricky little molecules and a lot of them are not doing what they are supposed to do. A helpful source is checking out the Human Protein Atlas (http://www.proteinatlas.org/). I prefer certain companies that do a lot of testing (Cell Signaling Technologies) but they still have bad apples in their catalogue. In the end, it is your risk. Some companies offer a "bonus" if you share your experience with the antibody (I think abcam does). Keep in mind that many antibodies are not as specific as you think. Positive controls are the key. If you have a lysate, cell type of tissue where you know your antigene should not be expressed, you have a nice negative control. With positive control, I mean, that the antibody should recognize a protein of a certain size, should locate in stainings to the right place (nucleus, membrane etc.), being responsive to siRNA treatments against the gene, and the very best: show the same pattern in all these assays as a second unrelated antibody against the same protein. There are so many bad stainings and antibodies out there, that it is very hard to figure out how to find a good one and make good stainings. Many people like to use the peptide against which the antibody was raised as a control. That is hypocrisy. If the antigen was bad or not specific enough, then the antibody still will bind lots of strange epitopes that have nothing to do with the actual protein of interest. Then many rabbit sera seem to be "contaminated" with anti-keratin antibodies and one can get nice epithelial stainings that have little to do with the antigen you are interested in.

As I mentioned, antibodies are tricky, expensive and not as straight forward as DNA and RNA based assays.

Great answers above. I agree with Pietro. It is a lot of work for companies to test all their antibodies in all applications and in all species/tissue types, however you should not have to pay to test the antibody for the company. A few companies will provide you with a free trial size to test in your specific application if they have not tested it already themselves. They usually request a review of the antibody once your test is complete, but it may be worth it to get the free trial.

Good luck!