Pretty formatting of pairwise alignment: XML/JSON to a multiline string

Hey all, I'm getting blastp results in a JSON, and want to turn them into a readable format to display on a website

to formulate this as a general question about pairwise alignment, I'm looking for a function (don't matter which programming language) that takes in the following arguments: Positions of the alignment on sequence_a (QUERY_FROM, QUERY_TO) Positions of the alignment on sequence_b (HIT_FROM, HIT_TO) The alignment length (ALIGN_LEN), sequence_a, sequence_b (QSEQ,HSEQ) Plus the middle line (MIDLINE)

and returns a long string with line breaks that is a readable representation of the alignment with positions on the sequences.

Sample input:

{
"query_from": 4,
"query_to": 96,
"hit_from": 1,
"hit_to": 99,
"align_len": 100,
"qseq": "MSDYSTMSSGYCSLEVELEDCFFTAK----RNLQSKQPTKNLCKAVEETWHPPTIQEIKQKIDSY---EKFCLGMKLSEDGYYTGFIKVVGLKLRRPVTV",
"hseq": "MTVDSSMSSGYCSLDEELEDCFFTAKTTFFRNLQSKQPSKNVCKAVEETQHPPTIQEIKQKIDSYNSREKHCLGMKLSEDGTYTGFIK-VHLKLRRPVTV",
"midline": "M+  S+MSSGYCSL+ ELEDCFFTAK    RNLQSKQP+KN+CKAVEET HPPTIQEIKQKIDSY   EK CLGMKLSEDG YTGFIK V LKLRRPVTV"
}

Requested output:

Query  4   MSDYSTMSSGYCSLEVELEDCFFTAK----RNLQSKQPTKNLCKAVEETWHPPTIQEIKQ  59
           M+  S+MSSGYCSL+ ELEDCFFTAK    RNLQSKQP+KN+CKAVEET HPPTIQEIKQ
Sbjct  1   MTVDSSMSSGYCSLDEELEDCFFTAKTTFFRNLQSKQPSKNVCKAVEETQHPPTIQEIKQ  60

Query  60  KIDSY---EKFCLGMKLSEDGYYTGFIKVVGLKLRRPVTV  96
           KIDSY   EK CLGMKLSEDG YTGFIK V LKLRRPVTV
Sbjct  61  KIDSYNSREKHCLGMKLSEDGTYTGFIK-VHLKLRRPVTV  99

Thanks!

formatting alignment blast

• 1,357 views • link

written 6.3 years ago by Yotam ConstantiniBiostar •

You can use blast_formatter program included in blast+ package, if you save the output in -outfmt 11 first.

$ blast_formatter -h
USAGE
  blast_formatter [-h] [-help] [-rid BLAST_RID] [-archive ArchiveFile]
    [-outfmt format] [-show_gis] [-num_descriptions int_value]
    [-num_alignments int_value] [-line_length line_length] [-html]
    [-sorthits sort_hits] [-sorthsps sort_hsps]
    [-max_target_seqs num_sequences] [-out output_file] [-parse_deflines]
    [-version]

Take a look at biopython blast parser as well.

• 0 views • link

written 6.3 years ago by GenoMax

Thank you for the quick reply. I need a solution for the input I've written, as I can only use the JSON data I'm getting. I'm looking for some actual source code I can rewrite into my application

• 0 views • link

written 6.3 years ago by Yotam ConstantiniBiostar •

1 answer

I ended up just writing it in PL/SQL, if you ever need to translate to another programming language, here it is:

function pretty_pairwise_alignment(
    ALIGN_LEN integer,
    query_from INTEGER,
    query_to INTEGER,
    hit_from INTEGER,
    hit_to INTEGER,
    qseq CLOB,
    hseq CLOB,
    midline  CLOB
) return clob is
    pretty_output clob;
    line_break varchar2(6) := chr(13)||chr(10);
    padding_constant integer := 4;
    alignment_consumed integer := 0;
    amount integer := 60;
    query_index integer := query_from;
    query_end integer;
    hit_index integer := hit_from;
    hit_end integer;
    sub_query varchar2(60);
    sub_hit varchar2(60);
    sub_midline varchar2(60);
begin
    dbms_lob.createtemporary(pretty_output,false);
    -- determin the amount of charachters in the stringified positions
    padding_constant := FLOOR(LOG(10,greatest(query_to,hit_to))) + 3;


    while (ALIGN_LEN > alignment_consumed) loop
        -- if this is the last line, fix the amount of charachter in this line
        if (ALIGN_LEN - alignment_consumed <= 60) then amount := ALIGN_LEN - alignment_consumed; end if;

        sub_query := DBMS_LOB.SUBSTR(qseq, amount, alignment_consumed+1);
        sub_midline := DBMS_LOB.SUBSTR(midline, amount, alignment_consumed+1);
        sub_hit := DBMS_LOB.SUBSTR(hseq, amount, alignment_consumed+1);
        query_end := query_index + amount - REGEXP_COUNT(sub_query,'-') - 1;
        hit_end := hit_index + amount - REGEXP_COUNT(sub_hit,'-') - 1;

        DBMS_LOB.APPEND(pretty_output,
            'Query  '||RPAD(TO_CHAR(query_index,'fm999999'),padding_constant)||
            sub_query||'  '||query_end||line_break||
            RPAD(' ',7+padding_constant)||sub_midline||line_break||
            'Sbjct  '||RPAD(TO_CHAR(hit_index,'fm999999'),padding_constant)||
            sub_hit||'  '||hit_end||line_break||line_break
        );

        query_index := query_end + 1;
        hit_index := hit_end + 1;
        alignment_consumed := alignment_consumed + amount;
    end loop;

    return pretty_output;
end pretty_pairwise_alignment;

• 0 views • link

written 6.3 years ago by Yotam ConstantiniBiostar •