WORKING EXAMPLE:
If boost libraries are installed (see REQUIREMENTS), then a working example can be immediately obtained by typing the following at the terminal:
$ cd example
$ python runExample.py
This will first produce an executable of Telescoper, for a fixed readlength, convert the split fastq/fasta data to Telescoper's format, then run Telescoper. Running should take less than 15 minutes (anything longer than a couple of minutes generally indicates that it is running successfully). A fasta file corresponding to each extended seed will be produced in output/fastas/.
REQUIREMENTS:
- Boost must be installed. If you don't have boost installed, you can either install it, or just download the libraries. If downloading them from http://www.boost.org/ then edit the makefile to include them in place of the path /Users/maayanbresler/bin/boost2/boost_1_51_0. If you do have boost installed, then there is no need to modify the makefile, and everything should compile without trouble.
- RAM should be at least 8 GB.
GENERAL STEPS:
See "example" directory for a working example (see "WORKING EXAMPLE" section above for a description).
1) Create a set of "seeds"-- these are strings of, say, 700 bp, that Telescoper will extend from.
2) Edit the makefile to reflect your read length for the short reads, by setting TELE_L=<readlength>.
3) Make: enter the "code" directory, and type "make". This will make a file Telescoper<readlength>, and preprocessData<readlength>.
4) Run "preprocessData<readlength> <readstorename> <reads1.fastq> <reads2.fastq> (fasta formatted input is also acceptable, since quality scores are not used at all in Telescoper's current version). <readstorename> is a prefix for a pair of files that will be produced, containing the reads.
5) Run "converter.py <fastq_1.fastq> <fastq_2.fastq> <readstorename2>" on the long-insert read data to convert it to expected format. (Fasta is also acceptable). Later versions of this software will use the preprocess command for the long-insert data as well. Note that the input read format is expected to be ---> <---- for the long insert data (same as for the short insert data).
6) Run telescoper<readlength> with options specified, as described in the manual. Note that the output directory should have a subdirectory named "fastas" into which the fasta files corresponding to the seeds will be printed.
NOTES / Future work / performance issues:
- This is a beta version of Telescoper, not tested on many real read data sets. So far, it serves largely as a proof of concept of the ideas, but work is under way to improve it. Still, it may perform well on certain data sets, especially when the coverage is high and the short-insert length is reasonably long.
- Telescoper's error correction algorithm is currently still a development version. Performance may improve by truncating reads to a shorter length before running telescoper on them, if the error rate is high. Reads beginning with an "N" should have that first base removed before running Telescoper-- it hashes reads according to their first 20 bases, so will miss them if there is an error in the beginning. If many long-insert reads are available, performance will improve by discarding ones that are low quality or short.
- Telescoper is not optimized for tandem repeats on a scale smaller than the long-insert distribution will resolve. Much better tandem-repeat handling is a priority for version 1.0.
- Output: Telescoper has some options to produce more detailed outputs. If the directory <outdir>/pathpiecegraphs/ exists, then Telescoper will print a txt file corresponding to the large extension graph.
