How to write bigWig when input comes from memory (e.g. 3 vectors) instead of a file?

[ghuls]

How to write bigWig when input comes from memory (e.g. 3 vectors) instead of a file?

Suppose I have 3 sets of vectors for each chromosome with chromosome names (all the same), starts and ends for each interval (already correctly sorted), how can I write them directly to a bigWig file?

let chroms = vec!["chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY"];
let starts = vec![530, 538, 584, 713, 751, 860, 865, 873, 879, 902];
let ends = vec![538, 584, 713, 751, 860, 865, 873, 879, 902, 923];

[jackh726]

So sorry! I never saw this. This is definitely doable, I'll share a code snippet tomorrow or this weekend.

[jackh726]

This works:

use std::collections::HashMap;
use bigtools::bed::bedparser::BedParser;
use bigtools::bedchromdata::BedParserStreamingIterator;
use bigtools::{BigWigWrite, Value};

let chroms = vec![
    "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY",
];
let starts: Vec<u32> = vec![530, 538, 584, 713, 751, 860, 865, 873, 879, 902];
let ends: Vec<u32> = vec![538, 584, 713, 751, 860, 865, 873, 879, 902, 923];

let iter = chroms
    .into_iter()
    .zip(starts.into_iter().zip(ends.into_iter()));
let iter = iter.map(|(chrom, (start, end))| {
    Ok::<(String, Value), std::io::Error>((
        chrom.to_string(),
        Value {
            start,
            end,
            value: 0.0,
        },
    ))
});

let vals_iter = BedParserStreamingIterator::new(BedParser::wrap_iter(iter), true);

let chrom_map = HashMap::from([("chrY".to_string(), 57_227_415)]);

let runtime = tokio::runtime::Builder::new_current_thread()
    .build()
    .expect("Unable to create runtime.");

let outb = BigWigWrite::create_file("test.bigWig".to_string());
outb.write_singlethreaded(chrom_map, vals_iter, runtime)
    .unwrap();

I've added it as a test too.

[jackh726]

Just pushed some changes to make this a bit nicer (don't have to call to_string() and don't need to Ok wrap):

let chroms = vec![
    "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY", "chrY",
];
let starts: Vec<u32> = vec![530, 538, 584, 713, 751, 860, 865, 873, 879, 902];
let ends: Vec<u32> = vec![538, 584, 713, 751, 860, 865, 873, 879, 902, 923];

let iter = chroms
    .into_iter()
    .zip(starts.into_iter().zip(ends.into_iter()));
let iter = iter.map(|(chrom, (start, end))| {
    (
        chrom,
        Value {
            start,
            end,
            value: 0.0,
        },
    )
});

let vals_iter = BedParserStreamingIterator::new(BedParser::wrap_infallible_iter(iter), true);

let chrom_map = HashMap::from([("chrY".to_string(), 57_227_415)]);

let runtime = tokio::runtime::Builder::new_current_thread()
    .build()
    .expect("Unable to create runtime.");

let outb = BigWigWrite::create_file("test.bigWig".to_string());
outb.write_singlethreaded(chrom_map, vals_iter, runtime)
    .unwrap();

This will be in the next release.