Error following matlab's working with genomes

I tried following matlab's exapmle of how to memmorry map a genome file ( https://www.mathworks.com/help/bioinfo/examples/working-with-whole-genome-data.html ), but am getting the foollowing error -

Array indices must be positive integers or logical values.

Error in nt2int (line 92)

seq = map((uint8(nt) + 1) - uint8('a'));

I am using 'hg38.fa' from the UCSC genome browser.

Thank you,

Guy

[Walter Roberson]

Give the command

dbstop if error

and run the code. When it stops, look at double(nt) . If there are any entries in it less than 97 then you would get the indexing problem.

In particular you would get the error if nt contains any capital letters.

[Guy Nir]

Thank you Walter.

I did as you suggested. Indeed, the lowest entry is 49.

Obviously, char(49)=1. So, I'm not sure yet whether the genome sequence contains numbers or that the code for some reason turned ot a chrracter into a number. All I basically did is to follow the example a the URL I gave.

Guy

[Walter Roberson]

I cannot seem to locate a copy of hg38.fa at the moment. The primary source is in Southern California, where there are quite a lot of fires at the moment, so it would not surprise me if they temporarily shut down the servers to make it easier to filter the air.

[Guy Nir]

This is where I downloaded ‘hg38.fa.gz’. After downloading I gunzipped it.

https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/

I’m not near my laptop so cannot verify it actually downloads.

Thank you, Guy

[Walter Roberson]

Unfortunately that is timing out for me today.

[Guy Nir]

Thank you for trying

Answer 1

[Guy Nir]

I added a command to remove any character which is not a letter:

% Remove numbers
    numIdx=find(~isletter(charData));
    charData(numIdx)='';
    % Convert to integers
    intData = nt2int(charData);

I think Matlab should consider adjusting some of their functions to whole-genome files (instead of whole chromosmes). Other algorithms use these numbers within the FASTA file for info about these sequences.