Extremely fast and accurate open modification spectral library searching of high-resolution mass spectra using feature hashing and graphics processing units

Abstract

Open modification searching (OMS) is a powerful search strategy to identify peptides with any type of modification. OMS works by using a very wide precursor mass window to allow modified spectra to match against their unmodified variants, after which the modification types can be inferred from the corresponding precursor mass differences. A disadvantage of this strategy, however, is the large computational cost, because each query spectrum has to be compared against a multitude of candidate peptides. We have previously introduced the ANN-SoLo tool for fast and accurate open spectral library searching. ANN-SoLo uses approximate nearest neighbor indexing to speed up OMS by selecting only a limited number of the most relevant library spectra to compare to an unknown query spectrum. Here we demonstrate how this candidate selection procedure can be further optimized using graphics processing units. Additionally, we introduce a feature hashing scheme to convert high-resolution spectra to low-dimensional vectors. Based on these algorithmic advances, along with low-level code optimizations, the new version of ANN-SoLo is up to an order of magnitude faster than its initial version. This makes it possible to efficiently perform open searches on a large scale to gain a deeper understanding about the protein modification landscape. We demonstrate the computational efficiency and identification performance of ANN-SoLo based on a large data set of the draft human proteome. ANN-SoLo is implemented in Python and C++. It is freely available under the Apache 2.0 license at https: //github.com/bittremieux/ANN-SoLo.

2018 YPIC Challenge: A Case Study in Characterizing an Unknown Protein Sample

“…As an alternative strategy to find PTMs we have employed spectral networking. 31 A spectral network was constructed by representing each consensus spectrum as a node in a graph and connecting two nodes if their corresponding spectra are highly similar as measured by the shifted dot product 20,32 (figure 3). Because the shifted dot product takes mass shifts induced by a modification into account while matching two spectra the spectral network will contain connections between modified peptides and their unmodified counterparts.…”

Evaluation of open search methods based on theoretical mass spectra comparison

“…More recently, MODPlus [16] and Open-pFind [15] select the set of candidate PSMs with sequence tags and then perform realignments. ANN-SoLo [10,11] and the Hybrid search [9] approximate the improved score obtained taking modified peaks into account (ANN-SoLo's shifted dot product, hybrid search's cosine similarity). On the other hand, some OMS methods still rely on Strategy1 (e.g.…”

“…Since 2015, several very rapid OMS methods have been developed. For a few of them, the reference spectra come from consensus templates already observed, identified and stored in spectral librairies [8][9][10][11]; for others, the reference spectra, so-called theoretical spectra in that case, are generated from a protein database by simulating an ideal fragmentation of peptides (see Fig. 1) [12][13][14].…”

ppx: Programmatic access to proteomics data repositories

“…As an example, we used ppx to help reanalyze a fractionated HEK293 cell line dataset [13] with the open modification spectral library search engine ANN-SoLo [24,30]. Our goal was to reproduce the analysis originally presented by Bittremieux et al [24], but with an updated version of ANN-SoLo.…”