fastafilt.c

   1 /*
   2  * fastafilt, filter fasta files based on its protein complexes
   3  *
   4  * Copyright (C) 2012 Ali Gholami Rudi <ali at rudi dot ir>
   5  *
   6  * This program is released under the modified BSD license.
   7  */
   8 #include <stdio.h>
   9 #include <string.h>
  10
  11 #define NPROTS          (1 << 12)
  12 #define NLEN            (128)
  13
  14 static char prots[NPROTS][NLEN];
  15 static int nprots;
  16
  17 static int prot_find(char *name)
  18 {
  19         int i;
  20         for (i = 0; i < nprots; i++)
  21                 if (!strcmp(name, prots[i]))
  22                         return i;
  23         return -1;
  24 }
  25
  26 static void prot_add(char *name)
  27 {
  28         if (prot_find(name) >= 0)
  29                 return;
  30         strcpy(prots[nprots++], name);
  31 }
  32
  33 static void prot_read(FILE *fin)
  34 {
  35         char name[NLEN];
  36         while (fscanf(fin, "%s", name) == 1)
  37                 prot_add(name);
  38 }
  39
  40 static void fasta_filt(FILE *fin)
  41 {
  42         char line[256];
  43         char name[NLEN];
  44         int matched = 0;
  45         while (fgets(line, sizeof(line), fin)) {
  46                 if (line[0] != '>') {
  47                         if (matched)
  48                                 printf("%s", line);
  49                         continue;
  50                 }
  51                 sscanf(line + 1, "%s", name);
  52                 matched = prot_find(name) >= 0;
  53                 if (matched)
  54                         printf("%s", line);
  55         }
  56 }
  57
  58 int main(int argc, char *argv[])
  59 {
  60         FILE *fin = argc > 1 ? fopen(argv[1], "r") : NULL;
  61         if (!fin) {
  62                 printf("usage: %s in.fasta <complexes >out.fasta\n", argv[0]);
  63                 return 0;
  64         }
  65         prot_read(stdin);
  66         fasta_filt(fin);
  67         fclose(fin);
  68         return 0;
  69 }