clj-deconstruct-sigs is a Clojure port of deconstructSigs.
A comprehensive description can be found here. In short, treating an individual's somatic mutation pattern as a combination of known distinct signatures could lead to insights of cancer's characteristics.
The main function is clj-deconstruct-sigs.core/which-signatures
, which takes a sample tumor and a reference signatures matrix and returns a map with the following keys.
:seed-idx - Index of the signature that was used as the initial seed
:weights - A map of weight indices to weight. :weights* with zero contribution entries
:weights* - A map of non-zero weight indices to weight. Note that entries that have zero contribution will be missing.
:product - matrix product of :weights and the provided signatures
:unknown - 1 minus the sum of weights
:diff - element wise difference of the provided sample tumor and the product
:error-sum - Square root of the sum of the element wise square of :diff. Indicates how close the :product is from the original input
Here is a repl session that adds more context.
(require '[clj-deconstruct-sigs.core :refer :all])
;; A 96 element vector representing the substitution patterns as per https://cancer.sanger.ac.uk/cosmic/signatures.
;; Start from A[C>A]A and ends with T[T>G]T
(def sample-tumor [0.00131420164040066,....])
;; Reference set of signatures to use, represented with m rows x 96 columns.
;; This library provides cosmic.signature-data.generated/latest-cosmic-signatures which is based on parsing the latest
;; signatures at https://cancer.sanger.ac.uk/cancergenome/assets/signatures_probabilities.txt
(require '[clj-deconstruct-sigs.data.cosmic])
(which-signatures sample-tumor clj-deconstruct-sigs.data.cosmic/cosmic-signatures)
=>
{:seed-idx 25,
:weights {7 0.36016636298023036,...},
:weights* {0 0.0, ...},
:product [6.690145741717268E-4,...],
:unknown 0.09329024976612943,
:diff [6.451870662289333E-4,...],
:error-sum 0.023682681269721964}
Current implementation uses clj-vectorz.
While faster than core.matrix
itself, it is dramatically slower than the original R implementation.
Help on tackling performance will be greatly appreciated!
Copyright 2019 Xcoo, Inc.
Licensed under the Apache License, Version 2.0.
Can you improve this documentation?Edit on GitHub
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close