Rodrigues NTL , Bland T , Borrego-Pinto J , Ng K , Hirani N , Gu Y , Foo S , Goehring NW .

Francis Crick Institute, FC001086 Cancer Research UK, FC001086 Medical Research Council , FC001086 Wellcome Trust , BB/T000481/1 Biotechnology and Biological Sciences Research Council , P40 OD010440 NIH HHS , 220790/Z/20/Z Wellcome Trust

	Fig. 1. Autofluorescence in C. elegans is correlated across a broad spectrum of wavelengths. (A) Normalized fluorescence intensity for embryos expressing PAR-6::GFP, PAR-3::GFP or LGL::GFP (strains: KK1248, KK1216 and NWG0285, respectively) relative to unlabeled N2 embryo controls reveals that AF contributes substantially to total measured fluorescence signal. Data for individual embryos are shown with mean indicated. (B) GFP channel images of unlabeled N2 and LGL-1::GFP embryos captured using identical parameters. Embryos are taken from dataset in A. (C) Subtraction of the fluorescence emission spectrum of unlabeled embryos (AF signal) from that of GFP::PAR-6 expressing embryos yields a calculated GFP spectrum that is indistinguishable from the expected theoretical GFP spectrum (purple dashed line) demonstrating AF and GFP signals are additive. Fluorescence emission from embryos was measured in 20 nm wavebands with midpoints from 520 to 700 nm. Data are mean±s.d. N2 (unlabeled), five embryos; PAR-6::GFP, five embryos. (D) Schematic of the SAIBR approach. Raw GFP channel images consist of both GFP and AF signal. True GFP signal is obtained by subtracting inferred AF signal in the GFP channel derived from AF measurement in a second channel. (E) Individual pixel values are well correlated between Gaussian-filtered (radius=1) AF and GFP channels, and are similar between embryos. Pixels from a region of interest comprising the entire embryo and a small section of surrounding background were used for the regression fit. A random selection comprising 10% of all pixels is shown color-coded per embryo. Histograms of intensity values for GFP and AF channels are shown for reference. (F) Comparison of per-pixel correlation with data obtained from whole embryo means. Lines indicate per-pixel regression from individual embryos as in E. An overlay of mean whole-embryo fluorescence values (circles) is shown below. (G) Application of SAIBR to a GFP::CDC-25.3-expressing embryo (DG4190). Subtraction of inferred AF signal reveals nuclear accumulation beginning at the start of the four-cell stage, and protein release at NEBD (t3). Scale bars: 10 µm.
	Fig. 2. Schematic summary of SAIBR workflow. The first step for two-channel SAIBR is to define the correction function ƒSAIBR. ƒSAIBR is defined by imaging multiple unlabeled calibration samples in which all fluorescence signal arises from autofluorescence, and performing a linear regression on fluorescence values in the GFP and AF channels. In step 2, to isolate GFP signal in GFP-labeled samples, measured AF channel signal is used to infer AF signal (AF*) in the GFP channel using ƒSAIBR, and inferred AF signal is then subtracted from the observed GFP channel signal to yield the ‘true GFP’ signal. In three-channel SAIBR, one must compensate for red fluorophore (e.g. RFP) signal bleedthrough into the AF channel. Therefore, for three-channel SAIBR, ƒSAIBR is defined by imaging samples expressing only RFP and performing a multiple linear regression to correlate observed AF and RFP channel signals to signal in the GFP channel. ƒSAIBR can then be used to infer AF in GFP-labeled samples based on AF and RFP channel signal, which can then be subtracted from the from observed GFP channel signal to obtain the true GFP signal.
	Fig. 3. Quantitative analysis of GFP concentrations by SAIBR. (A-D) Raw (top) and SAIBR-corrected (middle) midplane images of zygotes expressing the indicated GFP fusions imaged in the GFP channel as shown, along with an associated fluorescence linescan taken as indicated (bottom). SAIBR reduces GFP channel signal to background in unlabeled embryos (A) and reveals a prominent membrane-localized signal (arrows, B,C) that, before correction, is only obvious in PAR-6::GFP embryos. The prominent depletion of signal in the pronuclear/spindle regions in GFP channel images (white arrowheads) largely disappears in SAIBR images, suggesting it is due to local exclusion of AF signal. Strains are KK1216, KK1248 and NWG0285. (E-G) Averaged membrane profiles taken from either raw (top) or SAIBR-corrected (bottom) images of for LGL-1::GFP (E), PAR-3::GFP (F) and PAR-6::GFP (G) shown relative to unlabeled N2 controls. Plasma membrane position is defined as x=0 µm. Data are mean±s.d. Number of embryos is indicated (parentheses). Membrane peaks are strongly enhanced for each GFP fusion-expressing embryo, while N2 profiles are reduced to background in all datasets. (H) Quantitation of mean embryo GFP signal after correction by either mean AF subtraction (green) or SAIBR (blue) for embryos expressing PAR-6::GFP in single (PAR-6GFP/+) or double copy (PAR-6GFP/GFP), PAR-3::GFP (PAR-3GFP/GFP) or LGL-1::GFP (LGL-1GFP/GFP). Mean fluorescence values for individual embryos shown with group mean and coefficient of variation (c.o.v.) indicated. Strains used are N2, KK1248, KK1216 and NWG0285. Scale bar: 10 µm.
	Fig. 4. SAIBR achieves similar results to more specialized AF correction regimes. (A) Raw fluorescence signal using GFP illumination (GFP channel, ex488/em510-560) for unlabeled (N2) or PAR-3::mNG embryos (strain is NWG0189). Mean values are indicated. (B) As in A for mNG-specific illumination (mNG channel, ex514/em525-575). (C) Comparison of mean normalized mNG fluorescence under three AF correction regimes, as indicated. SAIBR and mNG-specific illumination yield similar reductions in the coefficient of variation (c.o.v.) relative to standard GFP illumination with mean AF subtraction. (D) Midplane images of excited embryos imaged using the GFP channel or mNG channel, and either left unprocessed (raw) or subjected to SAIBR, as indicated. PAR-3::mNG and unlabeled N2 are shown for comparison. Scale bars: 10 µm. (E) Plots of profiles taken perpendicularly across the plasma membrane (along the red line in embryos in D) highlight similar performance of SAIBR and use of the mNG channel in reducing AF. (F) Midplane images of a single LGL-1::GFP embryo captured via narrow (ex488/em504-512) or broad (ex488/em504-557) emission bands shown either uncorrected or corrected via SAIBR. (G) Same embryo as in F corrected for AF via spectral unmixing using either reference-free (blind) or reference-calibrated (calibrated) automatic component extraction. Images in F and G are individually normalized to 0.01% saturated pixels. (H,I) Plots of a 16 µm profile taken perpendicularly across the plasma membrane and averaged over a domain spanning the posterior 30% for LGL-1::GFP (H) and unlabeled (I) embryos. Fluorescence normalized to respective peak LGL-1::GFP signal. Plasma membrane position is defined as x=0. Regions outside the embryo, the cytoplasm and number references to the appropriate imaging conditions are indicated in the legend below. Data are mean±95% CI [computed by bootstrapping, n=5 (LGL::1GFP), n=3 (unlabeled)].
	Fig. 5. Spectral AF correction is compatible with dual color labeled samples. (A) Red fluorophore spillover increases signal in the AF channel for highly expressed proteins. GFP, AF and RFP channel signal shown for TH209 (mCh::PAR-2) or NWG0033 (mCh::MEX-5) embryos relative to unlabeled N2. The increase in the AF channel is significant for NWG0033 due to higher levels of mCherry expression. P values were determined using an unpaired two-tailed t-test. (B) mCherry spillover shifts the AF versus GFP correlation. Measured signal in Gaussian-filtered (radius=1) AF versus GFP channel for a single NWG0033 (mCh::MEX-5) embryo compared with wild-type N2 control, color-coded by RFP fluorescence (mCherry). Pixels from a region of interest comprising the entire embryo and a small section of surrounding background were specified, and a random selection comprising 10% of all pixels is shown. Dashed line indicates AF versus GFP channel correlation obtained from unlabeled N2 embryos. Inset shows residuals as a function of mCherry expression. (C) Multiple regression fit for three-channel SAIBR of AF signal in the GFP channel as a combined function of signal in the AF and RFP channels for an embryo expressing only mCh::MEX-5. The same 10% sample of pixels is shown. (D) Predicted versus measured AF signal in the GFP channel based on the fit in C. (E) Three-channel SAIBR correction compensates for red fluorophore expression. Comparison of total PAR-6::GFP signal in embryos co-expressing mCh::MEX5 (strain is NWG0119) subject to either two- or three-channel SAIBR compared with two-channel SAIBR applied to embryos expressing PAR-6::GFP alone. Overcorrection due to mCherry excitation by 488 nm illumination conditions results in underestimation of true GFP signal when using two-channel SAIBR. P values were determined using an unpaired two-tailed t-test. (F) Images/profiles showing elimination of AF overcorrection in MEX-5::mCherry-expressing embryos. Two-channel SAIBR using unlabeled N2 embryos for calibration results in oversubtraction that is visible as negative values (arrow) of NWG0033 (mCherry::MEX-5) and in reduced cytoplasmic signal in NWG0119 (mCh::MEX-5, PAR-6::GFP). Oversubtraction is eliminated with three-channel SAIBR using NWG0033 (mCh::MEX-5) for calibration. (G) Simultaneous correction for embryo AF and egg-shell fluorescence. Eggshell fluorescence is strongly visible in the red fluorescence channel. Use of only two-channel SAIBR (calibration using only GFP and AF channels) results in oversubtraction of the AF signal, visible as negative values in two-channel SAIBR images (arrows). This is eliminated by three-channel SAIBR that takes into account the RFP channel, improving signal of the LGL-1 posterior domain (strain: NWG0285). Arrowheads highlight regions of suppressed LGL-1 signal at the domain boundary due to eggshell-induced AF oversubtraction in two-channel SAIBR that does not occur in three-channel SAIBR. Scale bars: 10 µm
	Fig. 6. SAIBR suppresses autofluorescence background in late embryo and larval stages. (A-C) SAIBR reduces background in late stage C. elegans embryos. (A) Lateral view of a 1.5-fold embryo (calibration sample, strain BOX241) highlighting presence of substantial AF, particularly in the posterior near the nascent intestine, which is suppressed by SAIBR. (B) Posterior AF obscures membrane localization of LGL-1::GFP, which is revealed by SAIBR. Images on the bottom row show magnified views of the posterior region (LUT – fire). (C) SAIBR applied to a dual-labeled LGL-1::GFP, PAR-6::mCherry embryo reveals basolateral membrane localization of LGL-1. (Top, left) Schematic of PAR-6 (magenta) and LGL-1 (green) localization in the developing intestine of a 1.5-fold stage embryo. (Top, right) PAR-6::mCherry labeling the apical domain of the intestine. (Second and third rows) Paired uncorrected and SAIBR images shown alone or merged with PAR-6::mCherry visualized in the RFP channel. Images on the bottom row show magnified views of the posterior intestine. (B,C) Different planes of a single 1.5-fold stage embryo (strain NWG0290). In B,C, arrows highlight nuclear void volume surrounded by both fluorophore-specific and autofluorescence signals; arrowheads highlight resolved LGL-1 membrane signal after applying SAIBR. (D) AF correction of C. elegans L1 larva reveals LGL-1::GFP basolateral localization in the intestine. Still images of larval intestine expressing either PAR-6::mCherry alone (calibration sample, top row; strain BOX241) or LGL-1::GFP and PAR-6::mCherry (bottom row, NWG0290) shown for RFP (mCherry), GFP and SAIBR-corrected channels along with a merge of RFP/SAIBR channels. Arrowheads indicate gut granules in the GFP and SAIBR channels to highlight their overcorrection in the SAIBR channel. Scale bars: 10 µm.
	Fig. 7. SAIBR effectively reduces AF in other model systems. (A,B) Suppression of cortical granule AF in starfish oocytes. (A) Image of a starfish oocyte labeled with ODF2::GFP; rhodamine-tubulin is shown in xy and yz planes during meiosis II. The spindle (tubulin, red) is perpendicular to the cell cortex with the mother centriole (ODF2, cyan) positioned at the spindle pole near the cell cortex. In raw uncorrected images, yolk granules (cyan) dominate the signal in the GFP channel (left, raw). Yolk granule signal is strongly suppressed by SAIBR, leaving the ODF2-marked centriole as the strongest signal (right, SAIBR). Scale bars: 10 µm. (B) Scaled single channel insets (10× zoom) of the ODF2-marked mother centriole and neighboring yolk granule from A, highlighting specific suppression of yolk granule fluorescence relative to ODF2 (LUT – inferno). Scale bars: 1 µm. (C) Suppression of vacuolar AF in fission yeast cells. Raw (blue) and SAIBR-corrected (purple) midplane images of yeast cells from an unlabeled (left cell) or Nem1::mNG-expressing strain (middle and right cells) imaged in the GFP channel. Quantitation of fluorescence linescans taken across cells as indicated (see red dashed lines). SAIBR reduces non-specific AF signal (mostly emitted from vacuoles – highlighted by black arcs) to background in the unlabeled cell, and reveals a prominent well-resolved Nem1-specific signal at the cortical ER (see orange arrows – middle cell) and nuclear envelope (see green arrows – right cell). Scale bars: 5 µm.

Alberti, A single laser method for subtraction of cell autofluorescence in flow cytometry. 1987, Pubmed

Alberti, A single laser method for subtraction of cell autofluorescence in flow cytometry. 1987, Pubmed
An, SKN-1 links C. elegans mesendodermal specification to a conserved oxidative stress response. 2003, Pubmed
Arribere, Efficient marker-free recovery of custom genetic modifications with CRISPR/Cas9 in Caenorhabditis elegans. 2014, Pubmed
Baharlou, AFid: a tool for automated identification and exclusion of autofluorescent objects from microscopy images. 2021, Pubmed
Barkoulas, Robustness and epistasis in the C. elegans vulval signaling network revealed by pathway dosage modulation. 2013, Pubmed
Beatty, The C. elegans homolog of Drosophila Lethal giant larvae functions redundantly with PAR-2 to maintain polarity in the early embryo. 2010, Pubmed
Billinton, Seeing the wood through the trees: a review of techniques for distinguishing green fluorescent protein from endogenous autofluorescence. 2001, Pubmed
Borrego-Pinto, Live Imaging of Centriole Dynamics by Fluorescently Tagged Proteins in Starfish Oocyte Meiosis. 2016, Pubmed , Echinobase
Borrego-Pinto, Distinct mechanisms eliminate mother and daughter centrioles in meiosis of starfish oocytes. 2016, Pubmed , Echinobase
Castiglioni, Epidermal PAR-6 and PKC-3 are essential for larval development of C. elegans and organize non-centrosomal microtubules. 2020, Pubmed
Chen, Facile autofluorescence suppression enabling tracking of single viruses in live cells. 2019, Pubmed
Cowen, Pontamine sky blue: a counterstain for background autofluorescence in fluorescence and immunofluorescence histochemistry. 1985, Pubmed
Croce, Autofluorescence spectroscopy and imaging: a tool for biomedical research and diagnosis. 2014, Pubmed
Davis, Accurate detection of low levels of fluorescence emission in autofluorescent background: francisella-infected macrophage cells. 2010, Pubmed
Dickinson, CRISPR-Based Methods for Caenorhabditis elegans Genome Engineering. 2016, Pubmed
Dokshin, Robust Genome Editing with Short Single-Stranded and Long, Partially Single-Stranded DNA Donors in Caenorhabditis elegans. 2018, Pubmed
Gross, Guiding self-organized pattern formation in cell polarity establishment. 2019, Pubmed
Heppert, Comparative assessment of fluorescent proteins for in vivo imaging in an animal model system. 2016, Pubmed
Hermann, Genetic analysis of lysosomal trafficking in Caenorhabditis elegans. 2005, Pubmed
Hoege, LGL can partition the cortex of one-cell Caenorhabditis elegans embryos into two domains. 2010, Pubmed
Kumar, Feasibility of in vivo imaging of fluorescent proteins using lifetime contrast. 2009, Pubmed
Lang, The PAR proteins: from molecular circuits to dynamic self-stabilizing cell polarity. 2017, Pubmed
Laufer, Segregation of developmental potential in early embryos of Caenorhabditis elegans. 1980, Pubmed
Lichten, Unmixing of fluorescence spectra to resolve quantitative time-series measurements of gene expression in plate readers. 2014, Pubmed
Mansfield, Autofluorescence removal, multiplexing, and automated analysis methods for in-vivo fluorescence imaging. 2005, Pubmed
McRae, Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning. 2019, Pubmed
Neumann, Simple method for reduction of autofluorescence in fluorescence microscopy. 2002, Pubmed
Pang, Autofluorescence removal using a customized filter set. 2013, Pubmed
Pincus, Autofluorescence as a measure of senescence in C. elegans: look to red, not blue or green. 2016, Pubmed
Pintard, Mitotic Cell Division in Caenorhabditis elegans. 2019, Pubmed
Reich, Regulated Activation of the PAR Polarity Network Ensures a Timely and Specific Response to Spatial Cues. 2019, Pubmed
Rodriguez, aPKC Cycles between Functionally Distinct PAR Protein Assemblies to Drive Cell Polarity. 2017, Pubmed
Roederer, Compensation in flow cytometry. 2002, Pubmed
Roederer, Cell-by-cell autofluorescence correction for low signal-to-noise systems: application to epidermal growth factor endocytosis by 3T3 fibroblasts. 1986, Pubmed
Sallee, Apical PAR complex proteins protect against programmed epithelial assaults to create a continuous and functional intestinal lumen. 2021, Pubmed
Shaner, Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. 2004, Pubmed
Shcherbo, Far-red fluorescent tags for protein imaging in living tissues. 2009, Pubmed
Shi, In vivo analysis of recycling endosomes in Caenorhabditis elegans. 2015, Pubmed
Steinkamp, Dual-laser, differential fluorescence correction method for reducing cellular background autofluorescence. 1986, Pubmed
Stiernagle, Maintenance of C. elegans. 2006, Pubmed
Szöllösi, Autofluorescence correction for fluorescence in situ hybridization. 1995, Pubmed
Terasaki, Redistribution of cytoplasmic components during germinal vesicle breakdown in starfish oocytes. 1994, Pubmed , Echinobase
Teuscher, Overcoming Autofluorescence to Assess GFP Expression During Normal Physiology and Aging in Caenorhabditis elegans. 2018, Pubmed
Tsukamoto, LIN-41 and OMA Ribonucleoprotein Complexes Mediate a Translational Repression-to-Activation Switch Controlling Oocyte Meiotic Maturation and the Oocyte-to-Embryo Transition in Caenorhabditis elegans. 2017, Pubmed
Van de Lest, Elimination of autofluorescence in immunofluorescence microscopy with digital image processing. 1995, Pubmed
Zhang, Tissue morphogenesis: how multiple cells cooperate to generate a tissue. 2010, Pubmed
Zimmermann, Spectral imaging and its applications in live cell microscopy. 2003, Pubmed
Zipperlen, Roles for 147 embryonic lethal genes on C.elegans chromosome I identified by RNA interference and video microscopy. 2001, Pubmed