Welcome to fingerpuppet! 

What's a fingerpuppet?
fingerpuppet is a Python module (plus extra goodies) that is designed for the analysis of RNA-Seq data.

What's RNA-Seq?
Brother, if you don't know, you're living in the stone age. You may be unfamiliar with it because you're not a biologist, or more specifically you
aren't a geneticist, or even more specifically because you don't study global gene expression. I will assume for the purposes of this README that
you aren't a geneticist, or a biologist even, and that you probably failed biology in junior high because you got naseous when asked to dissect a
frog. So let's start from the beginning. You are not a single piece; rather, you are made up of billions of tiny balls called cells. All of the 
information that defines you biologically is encoded in molecules called DNA, with a copy of your DNA present in almost every cell. DNA forms a book
of you, written in a very specific language with four letters in its alphabet (A,T,G, and C). DNA is organized into units called genes; these are subsets of your
DNA that describe the structure of each one of the tiny pieces of protein that make up your cells, along with the structure of lots of other types
of molecules that your cells need. Just like DNA, proteins are made up of a small number of molecules which are strung together in a linear sequence;
three-letter words in the DNA alphabet encode a single letter in the protein alphabet. To make, for instance, a protein, the gene encoding that 
protein is copied from the DNA into a similar molecule called RNA. This copying is based on a fundamental property of DNA: A's stick to T's and 
G's stick to C's. This is called complementarity, and is most fundamentally embodied in the fact that your DNA is stored in two complementary copies 
which are stuck together to form a double-spiral. To copy a gene from DNA into RNA, the spiral is split at the gene, and a complementary sequence 
of RNA is assembled to match the DNA for the gene. The RNA is then processed in many complex ways, and then its final sequence is translated into the 
sequence of a protein. On a simplistic level, if your cell wants a lot of a particular protein, it makes lots of copies of the corresponding RNA. This
fact, and the complementarity of DNA/RNA, is what makes RNA-Seq work. In three steps: 1) you extract all the RNA from a leaf, or someone's hair, or whatever,
2) you determine the sequence of all these molecules of RNA, and 3) you compare those sequences to the known DNA sequnce (aligning), in order to determine
what was being produced in your sample. Importantly, this doesn't just tell your what proteins were being made, but also how much, because you can count
the number of times a particular RNA sequence showed up in your sample. Unfortunately, it isn't that simple, and there's a lot of guess work and heavy
statistical analysis that goes into making this work. The first step is up to the biologist, and trust me, getting RNA out of your co-worker's toe nails
is no easy task. Even once you've gotten the RNA, transforming it into something you can use in the RNA-Seq machine (a process called library prepartion)
is very complex and delicate. For the purposes of fingerpuppet, we will assume this has already been done, if not perfectly, at least as well as possible.
fingerpuppet does attempt to compensate for flaws in library preparation, but it isn't perfect. The second step is performed by a program called an aligner.
An example would be bowtie or maq. These programs use a variety of techniques to find the best match within a large DNA molecule for the many small RNA
molecules in a sample. It's really easy and no one should be impressed with such software; I mean, all you do is line up the DNA and the RNA, and you're done.
Right? Wrong. First off, DNA is really big, like 3 billion letters long for a human, and RNA from these experiments is really small, say 35-70 letters,
so you really are trying to find a needle in a haystack. Er, wait. My mistake. Not *a* needle in a haystack, more like 25 million needles in a haystack for
an average RNA-Seq experiment. And did I mention that there are lots of errors in the RNA sequence, which means that a given RNA molecule won't necessarily
line up perfectly to the DNA segment it came from? Long story short, alignment is really hard, and many people have spent a long time getting it to be
efficient and accurate. But for fingerpuppet, we will also assume this has been done. What fingerpuppet takes as input is a compressed file of alignments,
called a BAM file, which many aligners can produce, and which contains for each RNA molecule its sequence and where it aligned to in the DNA. This is where
our fun begins.

To be continued...