When I was a grad student, I modified an existing technique to measure the activity of an ion channel that resides on intracellular membranes. The major advance was that I extended the so-called “nuclear patch-clamp technique” from Xenopus nuclei to mammalian cells*. For several reasons that are unimportant, this was of fairly high interest to the field and immediately people tried to replicate my findings. It turns out that like many things in science, there are subtle aspects to the protocol that greatly increase the probability of success which do not translate well to words on a written protocol. I literally traveled to several labs around the country to show them with my own hands how to do this type of recording. This is now routine in the field, with a couple groups who are exceptionally good at this technique getting a lot of collaboration requests. The point is that if you can’t reproduce something, contact the group who published the work and they will likely help you in any way you can. I have certainly benefitted from this, as our most recent paper** required many email/phone calls to a couple of other groups who were much more competent in the assays we were trying to perform. One last thing, the literature is inherently self-correcting. The notion that you cannot publish contradictory or negative results is a fallacy***.