Two particle picking procedures for filamentous proteins: SPHIRE-crYOLO filament mode and SPHIRE-STRIPER

Abstract

Structure determination of filamentous molecular complexes involves the selection of filaments from cryo-EM micrographs. The automatic selection of helical specimens is particularly difficult and thus many challenging samples with issues such as contamination or aggregation are still manually picked. Here we present two approaches for selecting filamentous complexes: one uses a trained deep neural network to identify the filaments and is integrated in SPHIRE-crYOLO, the other one, called SPHIRE-STRIPER, is based on a classical line detection approach. The advantage of the crYOLO based procedure is that it accurately performs on very challenging data sets and selects filaments with high accuracy.Although STRIPER is less precise, the user benefits from less intervention, since in contrast to crYOLO, STRIPER does not require training. We evaluate the performance of both procedures on tobacco mosaic virus and filamentous F-actin data sets to demonstrate the robustness of each method.

Structural basis of actin filament assembly and aging

“… 53 ), contrast transfer function (CTF) estimations were performed with CTFFIND4.13 (ref. 54 ) and F-actin segments were picked using the filament picking procedure in SPHIRE_crYOLO 55 , 56 using a box distance of 40 pixels per 27.8 Å and a minimum number of six boxes per filament. The resulting particles were extracted in a 384 × 384 pixel box and further processed into the pipeline of helical SPHIRE v.1.4 (ref.…”

A structural basis for prion strain diversity

“…49). We then trained the deep learning package crYOLO 50 to pick ME7 filaments using 100 example micrographs, as previously reported 23 . We imported the coordinates into RELION and extracted images of prion rod segments with different box sizes (ranging from 1,024 to 384 pixels) to perform reference-free two-dimensional (2D) classifications.…”

PspA adopts an ESCRT-III-like fold and remodels bacterial membranes

“…A subset of the resulting micrographs was subjected to manual filament tracing using E2HELIXBOXER (Ludtke, 2016). Manually traced filaments were used for neural network training and subsequent automated tracing with crYOLO 1.5 (Wagner et al, 2019(Wagner et al, , 2020. This way, a total of 55,800 rods were detected.…”