Initializing a 4-D array

**gkoenig** · 03-01-2008

Hello to all,

I need help initializing a 4-D array that will hold all instances from [0,0,0,0] to [4,4,4,4] (only using 1-4). I need it for a comparison chart. What I am trying to do is scan a 5 million character string for each particular 4-mer that occurs and keep count. Any help would be great. Below is the code I have so far.

Code:

char t1, t2, t3, t4 ;
        int index ; 
        int tupleCount[4][4][4][4] ;

 for( index = 0 ; index < strlen(seqData) - 3 ; ++index )
            {
                t1 = seqData[index] ;
                t2 = seqData[index + 1] ;
                t3 = seqData[index + 2] ;
                t4 = seqData[index + 3] ;
                // count the different 4-mers 
            }

Sorry, my 4-D array should hold (1,1,1,1) to (4,4,4,4).

**vart** · 03-01-2008

from [0,0,0,0] to [4,4,4,4] (only using 1-4).

I do not get it
0-4 is 5 values
using 1-4 values you cannot get [0,0,0,0] vector

so what do you want to do?
how seqData is related to what you need?

to fill 4D array - use 4 loops one inside the other

Code:

for(i=0;i<4;i++)
   for(j=0;j<4;j++)
      for(k=0;k<4;k++)
         for(l=0;l<4;l++)
            tupleCount[i][j][k][l] = ...;

**dwks** · 03-01-2008

Sorry, my 4-D array should hold (1,1,1,1) to (4,4,4,4).

Not (0,0,0,0)? Hmm. In that case, you either waste space or subtract one every time you index the array.

Code:

for( index = 0 ; index < strlen(seqData) - 3 ; ++index )

Computing strlen() is quite expensive. You probably shouldn't do it in a loop. Consider calculating this value beforehand and storing it in a variable.

Now. Unfortunately, this code is broken.

Code:

                t1 = seqData[index] ;
                t2 = seqData[index + 1] ;
                t3 = seqData[index + 2] ;
                t4 = seqData[index + 3] ;

When you use seqData, you get a char ****. When you use seqData[x], you get a char ***. And so on, until you get a char when using seqData[x][y][z][a].

t1, t2, t3, and t4 are all chars. Therefore, you can't the sort of assignment you have.

I'm not really sure what you're trying to do here. Perhaps you could explain what a 4-mer is, for a start.

**gkoenig** · 03-02-2008

Sorry for the confusion; I was rushing when I posted. You are right. I should have said (0,0,0,0) to (3,3,3,3).

I guess it will be easier if I elucidate a little. I am trying to scan a sequence of DNA from the first letter to the last letter -3 for all the 4-mers that occur. I need to collect all the 4-mers so that I can see the total distribution. The array is to hold my count of each 4-mer as I progress through the sequence. I was also matching 0->a, 1->c, 2->g, an 3->t. I am not sure if this is a redundant step because when I print my distribution it will contain letters, not numbers. Hopefully this helps.

**xuftugulus** · 03-02-2008

Of course it does. If you name your array as tupleCount, and use a suggested method to map a,c,g,t to 0,1,2,3, which i will name acgt_to_int() then:

Code:

/* TODO: Verify that the seqData pointer points to enough data */
t1 = acgt_to_int(seqData);
t2 = acgt_to_int(seqData + 1);
t3 = acgt_to_int(seqData + 2);
t4 = acgt_to_int(seqData + 3);
/* TODO: Validate that t1..t4 are correctly in range */
tupleCount[t1][t2][t3][t4]++;

I leave the TODO tags to you.

**gkoenig** · 03-02-2008

Is there a simple c function to translate a,c,g,t to 0,1,2,3 ??

**xuftugulus** · 03-02-2008

That is up to you to create. Using a function will mask the gory details of your conversion.
You can do something like:

Code:

int acgt_to_int(char *x)
{
    if(*x == 'a')
        return 0;
    ....
}

Guess the rest. Also note the fact that you should handle case-sensitivity, and find a return value to indicate error.
I'd suggest -1.

**gkoenig** · 03-02-2008

I am not sure what I did wrong but, I am getting all kinds of errors and warnings. I've looked through the code and can't pick up on anything. Any help would be great.

Code:

// A Program to count the count of 4-mers in a nucleotide sequence.

#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include <stdlib.h>

FILE *input ; 
        FILE *outpur ;
        char buffer[1000] ;        
        int i = 0 ;
        int c ;
        int w,x,y,z ;
        char tr ;
         
        char *seqData ;
       
        seqData = (char *)malloc(10000000) ;
        
       
        char t1, t2, t3, t4 ;
        int index ; 

int acgt_to_0123(char *tr)
    {
        if(*tr == 'a')
            {
            return 0 ;
            }
        else if(*tr == 'c')
            {
            return 1 ;
            }
        else if(*tr == 'g')
            {
            return 2 ;
            }
        else if(*tr == 't')
            {
            return 3 ;
            }
        else
            {
            printf("Non acgt character\n") ;
            exit(0) ;
            }
    }

main( int argc, char **argv )
    {
                
        
        // Initialize 4-D array
        int tupleCount[4][4][4][4] ;
        for(w = 0 ; w < 4 ; w++ )
            for(x = 0 ; x < 4 ; x++ )
                for(y = 0 ; y < 4 ; y++ ) 
                    for(z = 0 ; z < 4 ; z++ )
                        {
                            tupleCount[w][x][y][z] ;
                        }
        
        // Open input file to read from        
        if( ! ( input = fopen( argv[1], "r" ) ) )
            { 
                printf( "COULD NOT OPEN FILE %s - Exit!\n", argv[1]) ; 
                exit(1) ; 
            }        
        
        
        // Collect sequence from GenBank file
        while(fgets(buffer, 1000, input))
                        {
                        // start obtaining bases after ORIGIN
                        if(strstr(buffer, "ORIGIN")) 
                            {                                                  
                               while((c=getc(input)) != '/' && c != EOF)
                                    {
                                    if(c >= 'a' && c <= 'z')
                                        {
                                        seqData[i++] = c ;
                                        }
                                    }                            
                            }           
                        }
                        
        
        
        
        // Scan DNA sequence for each 4-mer 
        for( index = 0 ; index < strlen(seqData) - 3 ; ++index )
            {
                t1 = acgt_to_0123(seqData[index]) ;
                t2 = acgt_to_0123(seqData[index + 1]) ;
                t3 = acgt_to_0123(seqData[index + 2]) ;
                t4 = acgt_to_0123(seqData[index + 3]) ;
                
                
                // Accumulate a count to find distribution
                tupleCount[t1][t2][t3][t4]++ ;
                
            }
    
                   
        fclose(input) ;
    
        printf("Here is the distribution of 4-mers:\n\n%s", tupleCount ) ;     
        free(seqData) ; 
        return(0) ;
    }

thanks

**xuftugulus** · 03-02-2008

Code:

        char tr ;

Unused variable. Remove it.

Code:

        seqData = (char *)malloc(10000000) ;

It is not allowed. Global variables must be initialized in block scope ( inside main() ).
Why so many globals? Move those declarations to the start of the main() block.

Code:

        // Initialize 4-D array
        int tupleCount[4][4][4][4] ;
        for(w = 0 ; w < 4 ; w++ )
            for(x = 0 ; x < 4 ; x++ )
                for(y = 0 ; y < 4 ; y++ ) 
                    for(z = 0 ; z < 4 ; z++ )
                        {
                            tupleCount[w][x][y][z] ;
                        }

Exactly what is that supposed to mean? Forgot the =0 maybe?

Code:

        // Scan DNA sequence for each 4-mer 
        for( index = 0 ; index < strlen(seqData) - 3 ; ++index )
            {
                t1 = acgt_to_0123(seqData[index]) ;
                t2 = acgt_to_0123(seqData[index + 1]) ;
                t3 = acgt_to_0123(seqData[index + 2]) ;
                t4 = acgt_to_0123(seqData[index + 3]) ;
                
                
                // Accumulate a count to find distribution
                tupleCount[t1][t2][t3][t4]++ ;
                
            }

In the definition of your function acgt_to_0123, i proposed a pointer to character to be passed.
You instead pass the character itself. Change the function to work with characters as it seems to be more straightforward to you. (hint: remove all * from acgt_to_0123).

And a sidenote on correctness. Looking at the for loop in your code and read aloud what is going on. You are moving the index, ONE character at a time, but i presume that the 4-mers are 4 characters long? Or is it considered for your excercise valid to do things like:
YOUR INPUT: 'acaactgagatc'
YOUR 4-MERS: 'acaa', 'caac', 'aact', 'actg', 'ctga', 'tgag', 'gaga', 'agat', 'gatc'

**gkoenig** · 03-02-2008

Thanks for the input; will implement. What out said for the 4-mer is exactly right. The input sequence is 5 million bases and all of those examples I must collect. For instance:

count of 4-mers:

aaaa = 3
aaac = 2
.
.
.
tttt = 7

thats what i need printed out at the end.

Again, thanks for the help. I really appreciate it. When I am done this semester I will make sure to go back and get more fluent in the language.

**gkoenig** · 03-02-2008

Okay, I've gotten it to compile but now it will not stop running. Did I enter an endless loop somewhere?

**xuftugulus** · 03-02-2008

Exactly how big is the file being scanned? Post bytes. Because the call to strlen on every iteration of the for loop that does statistics, will do say if N is the size of the file N^2 memory accesses. It should finish, but if N is 1E+6, then after 1E+12 memory accesses. And i think N is bigger in your case. Compute the result of strlen(seqData) in a temporary variable. And save yourself the quadratic performance over N to linear. What an optimization !

**gkoenig** · 03-02-2008

Code:

int lengthDNA = strlen(seqData) ;
        
        // Scan DNA sequence for each 4-mer 
        for( index = 0 ; index < (lengthDNA - 3) ; ++index )
            {
                t1 = acgt_to_0123(seqData[index]) ;
                t2 = acgt_to_0123(seqData[index + 1]) ;
                t3 = acgt_to_0123(seqData[index + 2]) ;
                t4 = acgt_to_0123(seqData[index + 3]) ;
                
                
                // Accumulate a count to find distribution
                tupleCount[t1][t2][t3][t4]++ ;                
            }

The file I used as input was 11MB and the length of the sequence is 5 Million chars. Still keeps running endlessly

**xuftugulus** · 03-02-2008

Do a verification printf on lengthDNA, before the loop.

**gkoenig** · 03-02-2008

Code:

 int lengthDNA = strlen(seqData) ;
        
        printf("here is the lenght of the DNA: %d", lengthDNA) ;
        
        // Scan DNA sequence for each 4-mer 
        for( index = 0 ; index < (lengthDNA - 3) ; ++index )
            {
                t1 = acgt_to_0123(seqData[index]) ;
                t2 = acgt_to_0123(seqData[index + 1]) ;
                t3 = acgt_to_0123(seqData[index + 2]) ;
                t4 = acgt_to_0123(seqData[index + 3]) ;
                
                
                // Accumulate a count to find distribution
                tupleCount[t1][t2][t3][t4]++ ;                
            }

It stopped the loop and printed out the sentences. Why was that able to stop the loop?
Also, it did not print out the distribution of 4-mers.

Code:

printf("Here is the distribution of 4-mers:\n\n%s", tupleCount ) ;

Thread: Initializing a 4-D array

Thread Tools

Search Thread

Display

Initializing a 4-D array

Similar Threads

allocation and reallocation of memory dynamically of an integer array

Initializing a 2D Array in C

question about multidimensional arrays

Type and nontype parameters w/overloading

Initializing char * array