Take Home Exam

% Take home exam from my 201 Class
% removed any identifying features
% similar to DNA Strands to Amino Acids Paste

%%

% opening file
seq_file = fopen('nonnatural.txt'); % Opens the text file
temp5_3 = fscanf(seq_file,'%s'); % scanning file and saving it as a variable

% starting replication
A = {'A' 'T'
     'C' 'G'
     'G' 'C'
     'T' 'A'}; % creates matrix to for comparison of replication

for i = 1:length(temp5_3) % makes the for loop until end of the template

    row_A = strcmp(temp5_3(i),A(:,1)); % compares the template in 5'-3' to the matrix and saves it to variable
    synth5_3(i) = char(A(row_A,2)); %#ok<SAGROW> % chooses the corresponding DNA base and saves it to a synth variable as a character

end

% starts transcription

sense5_3 = synth5_3; % moves synth strand to sense strand in the same direction

B = {'A' 'U'
     'C' 'G'
     'G' 'C'
     'T' 'A'};

for i = 1:length(temp5_3)

    row_B = strcmp(sense5_3(i),B(:,1)); % compares the sense strand to the matrix and saves as a variable
    RNA5_3(i) = char(B(row_B,2)); % chooses the corresponding RNA base and saves it in the RNA variable

end

methionine_positions=strfind(RNA5_3,'AUG'); % finds all the start codons

C = {'UUU' 'F'
     'UUC' 'F'
     'UUA' 'L'
     'UUG' 'L'
     'CUU' 'L'
     'CUC' 'L'
     'CUA' 'L'
     'CUG' 'L'
     'AUU' 'I'
     'AUC' 'I'
     'AUA' 'I'
     'AUG' 'M'
     'GUU' 'V'
     'GUC' 'V'
     'GUA' 'V'
     'GUG' 'V'
     'UCU' 'S'
     'UCC' 'S'
     'UCA' 'S'
     'UCG' 'S'
     'CCU' 'P'
     'CCC' 'P'
     'CCA' 'P'
     'CCG' 'P'
     'ACU' 'T'
     'ACC' 'T'
     'ACA' 'T'
     'ACG' 'T'
     'GCU' 'A'
     'GCC' 'A'
     'GCA' 'A'
     'GCG' 'A'
     'UAU' 'Y'
     'UAC' 'Y'
     'UAA' '*'
     'UAG' '*'
     'CAU' 'H'
     'CAC' 'H'
     'CAA' 'Q'
     'CAG' 'Q'
     'AAU' 'N'
     'AAC' 'N'
     'AAA' 'K'
     'AAG' 'K'
     'GAU' 'D'
     'GAC' 'D'
     'GAA' 'E'
     'GAG' 'E'
     'UGU' 'C'
     'UGC' 'C'
     'UGA' 'X' % new amino acid
     'UGG' 'W'
     'CGU' 'R'
     'CGC' 'R'
     'CGA' 'R'
     'CGG' 'R'
     'AGU' 'S'
     'AGC' 'S'
     'AGA' 'R'
     'AGG' 'R'
     'GGU' 'G'
     'GGC' 'G'
     'GGA' 'G'
     'GGG' 'G'}; % Matrix for identifying amino acids

AA_seq = NaN; % preallocating the vector to not a number
ORF = RNA5_3(methionine_positions(2):end); % subsets the RNA starting with the first AUG
AA_pos = 1; % the first Amino Acid position in the character vector
ORF_pos = 1:3; % makes sure groups of 3 will be scanned

while ~strcmp(char(AA_seq(end)),'*'); % only wants the loop run until it hits a stop codon

    row_C = strcmp(ORF(ORF_pos),C(:,1)); % compares the 3 RNA bases to the matrix for its amino acid and saves it as a vector
    AA_seq(AA_pos) = char(C(row_C,2)); % chooses the corresponding amino acid and saves it in the Amino Acid sequence vector

    AA_pos = AA_pos + 1; % moves to the next Amino acid index in the vector
    ORF_pos = ORF_pos + 3; % moves to the next 3 bases to compare
end

AA_seq = char(AA_seq); % changes to a character vector

AA_org = {'MRCVIGKVAHDDVRK*'}; % orginal amino acid sequence
AA_new = {'MRCVIGKVAHDDVRKXLSSVLCACVDSKYGTK*'} % new amino acid sequence
org_value = strlength(AA_org); % length of the orginal strand
new_value = strlength(AA_new); % length of the new strand

if org_value > new_value % if the orginal strand is longer
    fprintf('orginal strand is longer') % prints the text
elseif new_value > org_value % if new strand is longer
    fprintf('new strand is longer') % prints the text
else % if strands are the same size
    fprintf('strands are the same size') % prints the text
end