Repository navigation
Running
The following file format is expected to execute the algorithm:
Each line represent a non TIS mRNA sequence, the sequence must contain 40 RNA bases (a, c, g or t). Beside the mRNA bases the line also contains three binary flags:
- True if the sequence has an AUG in phase.
- True if the sequence has a stop condon in the next 100 nucleotides in the downstream region.
- True if the sequence has the condon GAG in phase in the next 100 nucleotides in the downstream region.
Each codon is coded in a binary sequence following the table bellow:
| codon | binary representation |
|---|---|
| A | 1000 |
| C | 0100 |
| G | 0010 |
| T | 0001 |
Each binary flag is represented by the following code sequence:
| Flag Value | binary representation |
|---|---|
| True | 0001 |
| False | 0010 |
An exemple of one line of some input file:
0010010000100001100001000001010001000001100000011000100000010010010010001000100010000100100000100100000100010100000110000100000100101000000110000010000100010001000100010010
This repository contains seven organisms classes inputs files, each organism contains 10 inputs files. Those 10 files represent all non negatives TIS sequence for that organism, excepted for 1/10 of them that is used for the cross validation.
The expected number of clusters is the number of existence TIS sequences for the organisms.
The fold is divide by ten, but they are not multiple of ten, so the last folder may contain less sequence, and expected a different number of clusters.
Those organisms are detailed in the table bellow.
| organism | sequences per file | number of clusters | number of clusters for the last file | |:-:|:-:|:-:|:-:|:-:| | Rattus novergicus | 301 | 28 | 27 | Gallus gallus | 482 | 12 | 9 | Mus musculus | 5631 | 243 | 234 | Homo sapiens | 318246 | 14261 | 14256 | Drosophila melanogaster | 392306 | 17804 | 17802 | Caenorhabditis elegans | 404316 | 7887 | 7884 | Arabidopsis thaliana | 529100 | 21096 | 21897
Because those files are too larges, they are compressed using gzip. You can extract all include mentioned above by running:
make extract_input
This command will extract the files to the following paths: input, input-mpi and input-cuda
To run the k-modes sequencial, you should use the kmodes executable locale in the bin directory and the input locale in the input directory. Ex:
/bin/kmodes input/gallus/cluster1 12
Where the first argument is the input file, and the second is the number of clusters.
To run the k-modes sequencial, you should use the kmodes executable locale in the bin directory and the input locale in the input directory. Ex:
/bin/kmodes input/gallus/cluster1 12
Where the first argument is the input file, and the second is the number of clusters.
/bin/cuda-kmodes input-cuda/gallus/cluster1 12
Where the first argument is the input file, and the second is the number of clusters.