Antibodies were as follows: ATG7 (Cell Signaling Technology, 8558s; I want: AB_10831194; Diluction 1: 1,000), FIP200 (Cell Signaling Technology, 12436; Irrid: AB_2797913; Diluction 1: 1,000), LC3B (MBL International, M186-3; AB_10897859; Dilution 1: 1,000), ULK1 (Cell Signaling Technology 8054; I want: AB_11178668; Dilution 1: 1,000), Phospho-Ulk1 (SER757) (Cell Signaling Technology 9644; I Want: AB_2097841; Dilution 1: 1,000), Phospho-4BP1 (Thr37/46) (Cell Signaling Technology 2855; Irrid: AB_560835; Diluction 1: 1,000), Tex264（Sigma，HPA017739; Irrid：Irr.AB_1857910; 1：1,000），HSP90（ProteIntech 60318; Irrid; Irrid：ab_2881429; ab_2881429;稀释1：1,000; calcoco1），Calcoco1（Abclonal A7987; rid：rid：ab_2768684; diming：rid a ab_2768684; 9091; irrid：AB_2687579; RID：AB_628110; 10,000），山羊抗兔IgG，HRP链接IgG（细胞信号技术7074P2，我想要：AB_2099233稀释1：10,000），山羊抗兔IgG IgG HRP HRP ConJUGATE 1706516; irrid：AB_11125547; （A-21244;我想要：AB_2535812）。

　　The following were used: FluoroBrite Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher, A1896701), benzonase nuclease HC (Millipore, 71205-3), urea (Sigma, catalogue number U5378), sodium dodecyl sulfate (Bio-Rad, catalogue number 1610302), high-glucose and high-pyruvateDMEM（Gibco/Invitrogen，11995），无氨基酸的低葡萄糖DMEM（美国生物学，D9800-13），TCEP（Gold Biotechnology），瓜霉素（金生物技术，p-600-100）4906845001），胰蛋白酶（Promega，v511c），LYSC（Wako Chemicals，129-02541），EPPS（Sigma-Aldrich，目录号E9502），2-氯乙酰酰胺（2-氯乙酰酰胺）（Sigma-Aldrich，Sigma-Aldrich，cataL，c0267），catal and tmo catly 6 prount inter inter intectect and cate and catter intectect and tmt retectect and tmtect（cate）（catly 6）A34807），TMTPRO 16plex标签试剂（Thermo Fisher，目录号A44520），羟胺溶液（Sigma Catalog Number 438227），Empore SPE Disks C18（3M-Sigma-Aldrich-Aldrich Catalog编号66883-U），sep-p-c1883-cartridde（sep-cartridde）（WAT054960和WAT054925），Sola HRP SPE弹药筒，10毫克（Thermo Fisher，目录编号60109-001），高ph相反的相位肽分馏试剂盒（Thermo Fisher（Thermo Fisher）（Thermo Fisher），目录编号84868），Bio-Rad蛋白质蛋白质ige 000和cate 000 000 000 cate 000 cate 006（Sigma-Aldrich Cataloge编号E3024）。

　　HEK293（人类胚胎肾脏，胎儿，ATCC CRL-1573，RRID：CVCL_0045）和HELA（宫颈癌细胞系CCL-2； RRID：CVCL_0030）细胞在高葡萄糖和高丙酮酸DMEM DMEM中生长，并在10％fetal Fetal Calf Calf Calum和3％的co中均在5％的co中添加了5％的fetal dmem dmem。供应商提供了细胞系认证，HEK293细胞（来自ATCC）的核分型（GTG带核型）也由Brigham和妇女医院的Cytegenomics Core实验室进行。细胞保持在 <80% confluency throughout the course of experiments. HeLa cells lacking MAP1LC3 (ΔLC3) or GABARAP (ΔRAP) proteins were from a previous study19. Culture of human ES cells or iNeurons was carried out as described at https://doi.org/10.17504/protocols.io.br9em93e. In brief, human ES cells (H9, WiCell Institute) with TRE3G-NGN2 integrated into the AAVS site have been previously described32 and were cultured in E8 medium on Matrigel-coated plates. To generate iNeurons (i3-neurons) from ES cells, cells were plated at 2 × 105 cells per millilitre on day 0 on plates coated with Matrigel in ND1 medium (DMEM/F12, 1× N2 (Thermo Fisher), human brain-derived neurotrophic factor (10 ng ml−1, PeproTech)), human neurotrophin-3 NT3 (10 ng ml−1, PeproTech), 1× nonessential amino acids, human laminin (0.2 μg ml−1) and doxycycline (2 μg ml−1). The medium was replaced with ND1 the next day. The next day, the medium was replaced with ND2 neurobasal medium, 1× B27, 1× Glutamax, brain-derived neurotrophic factor (10 ng ml−1), NT3 (10 ng ml−1) and doxycycline (2 μg ml−1). On days 4 and 6, 50% of the medium was changed with fresh ND2. On day 7, cells were replated at 4 × 105 cells per well in ND2 medium supplemented with Y27632 (rock inhibitor; 10 μM). The medium was replaced the next day with fresh ND2 and on day 10 onwards 50% medium change was carried out until the experimental day (day 14 of differentiation unless otherwise noted). Pluripotency and neurogenesis markers exhibited the expected changes for all genotypes (Extended Data Fig. 12c) and visual inspection demonstrated the expected pattern of axons and dendrites for all genotypes.

　　Cells were plated in 10-cm or 15-cm, 6-well dishes the night before nutrient stress. DMEM was removed and cells were washed three times with DPBS and then resuspended in EBSS or DMEM lacking amino acids prepared as described previously5 (and in https://doi.org/10.17504/protocols.io.yxmvm32nbl3p/v1). For whole-cell proteomics experiments, cells were resuspended in EBSS or medium lacking amino acids as described previously7 for 12–18 h. For APEX2 proximity labelling and imaging experiments, cells were resuspended in EBSS + BafA1 (100 nM) for 3–4 h in the presence or absence of the indicated inhibitors.

　　YIPF4-, FIP200- and ATG7-knockout in HEK293 cells and ATG7-, YIPF4- and CALCOCO1-knockout in HeLa cell lines were carried out by plasmid-based transfection of Cas9–gRNA using the pX459 plasmid as described previously35 and at https://doi.org/10.17504/protocols.io.6qpvr3462vmk/v1. The following gRNAs, designed using the CHOPCHOP website (http://chopchop.cbu.uib.no/), were used: YIPF4: 5′-ATCTCGCGGCGACTCCCAAC-3′ and 5′-CGGCCTATGCCCCCACTAAC-3′; FIP200: 5′-ACTACGATTGACACTAAAGA-3′; ATG7 HEK293: 5′-ATCCAAGGCACTACTAAAAG-3′; CALCOCO1: 5′-AAGTTGACTCCACCACGGGA-3′ and 5′-CTAAGCCGGGCACCATCCCG-3′; YIPF3: 5′-CCATTTCGGGCGCCGCCCGC-3′ and 5′-GGCGGCGCCCGAAATGGAGC-3′. Puromycin selection was carried out 24–48 h after the transfection. Cells were given a day to recover from puromycin selection, and then single cells were sorted into a 96-well plate using fluorescence-activated cell sorting (FACS) on a SONY SH800S sorter. Individual clones were screened for deletion of the relevant gene by immunoblotting cell extracts with antibodies specific to the designed gene product. For N-terminal tagging of the YIPF4 locus, the gRNA 5′-TCGCCGCGAGATGCAGCCTC-3′ was cloned into pX459 and co-transfected with a repair template containing an mNEON Green cassette flanked by homology arms (pSMART-mNEON-YIPF4) into HEK293 and HEK293 FIP200−/− cells using Lipofectamine 3000 (as described at https://dx.doi.org/10.17504/protocols.io.5jyl8pj9dg2w/v1). After 7 days, a population of cells for both genotypes was sorted for the same level of mNEON Green signal. For deletion of YIPF4 in human ES cells (H9), gRNA (5′-AAGAGGTTATGGCTGGCTTC-3′) was ordered from Synthego. A 0.6 μg quantity of sgRNA was incubated with 3 μg SpCas9 protein for 10 min at room temperature and electroporated into 2 × 105 H9 cells using the Neon transfection system (Thermo Fisher) as described at https://doi.org/10.17504/protocols.io.rm7vzxy44gx1/v1. Out-of-frame deletions were verified by DNA sequencing with Illumina MiSeq and by immunoblotting. All cell lines were demonstrated to be mycoplasma negative.

　　A protocol for cell lysis and immunoblotting can be found at: https://doi.org/10.17504/protocols.io.4r3l226e4l1y/v1. Cells were cultured in the presence of the corresponding stress to 60–80% confluency in 10-cm or 15-cm, 6-well dishes. After the medium was removed, the cells were washed with DPBS three times. To lyse cells, urea buffer (8 M urea, 50 mM Tris pH 7.5, 150 mM NaCl, containing mammalian protease inhibitor cocktail (Sigma), PhosSTOP, and 20 units per millilitre of Benzonase (Millipore)) was added directly onto the cells. Cell lysates were collected by cell scrapers and sonicated on ice for 10 s at level 5, and lysates were cleared by centrifugation (15,000 r.p.m., 10 min at 4 °C). The concentration of the supernatant was measured by the BCA assay. For immunoblotting, the whole-cell lysate was denatured by the addition of LDS sample buffer supplemented with 100 mM dithiothreitol (DTT), followed by boiling at 95 °C for 5 min. A 10–20 μg quantity of each lysate was loaded onto a 4–20% Tris-Glycine gel (Thermo Fisher) or a 4–12% NuPAGE Bis-Tris gel (Thermo Fisher), followed by SDS–PAGE with Tris-glycine SDS running buffer (Thermo Fisher) or MOPS SDS running buffer (Thermo Fisher), respectively. For chemiluminescence western blots, the proteins were electro-transferred to PVDF membranes (0.45 µm, Millipore), and then the total protein was stained using Ponceau (Thermo Fisher). The membrane was then blocked with 5% non-fat milk (room temperature, 60 min) incubated with the indicated primary antibodies (4 °C, overnight), washed three times with TBST (total 30 min), and further incubated with either HRP-conjugated anti-rabbit or anti-mouse secondaries (1:5,000) for 1 h. After a thorough wash with TBST for 30 min, membranes were treated with Lightning Plus Chemiluminescence Reagent (PerkinElmer, NEL104001EA) after mixing the Enhanced Luminol Reagent and the Oxidizing Reagent 1:1. Mixed Chemiluminescence Reagent was added to the blot and incubated with gentle rocking for 1 min before imaging of the blot using the Bio-Rad ChemiDoc Imaging System. For the LI-COR western blots, the proteins were electro-transferred to nitrocellulose membranes and then the total protein was stained using Ponceau (Thermo Fisher). The membrane was then blocked with LI-COR blocking buffer at room temperature for 1 h. Then membranes were incubated with the indicated primary antibodies (4 °C, overnight), washed three times with TBST (total 30 min), and further incubated with either fluorescent IRDye 680RD goat anti-Mouse IgG H+L, or IRDye 800CW goat anti-rabbit IgG H+L secondary antibody (1:10,000) at room temperature for 1 h. After a thorough wash with TBST for 30 min, the near-infrared signal was detected using an OdysseyCLx imager and quantified using ImageStudioLite (LI-COR).

　　Detailed protocols can be found at https://doi.org/10.17504/protocols.io.8epv5xj9ng1b/v1. Double-knockout (YIPF3−/−YIPF4−/−) HEK293 cells were reconstituted with mCherry–YIPF4 (WT or LIR mutant) and GFP–YIPF3 (WT or LIR mutant) constructs and sorted for equal expression levels. Immunofluorescence was used to confirm proper localization of both YIPF3 and YIPF4. Then cells were plated on 10-cm plates and grown to 70% confluency. Cells were left untreated or starved using amino acid withdrawal for 2 h in the presence of BafA1 (100 nM). Cells were washed twice with cold PBS and then lysed in 0.8 ml NP-40 lysis buffer (100 mM Tris pH 7.4, 150 mM KCl, 0.1% NP-40, 0.5 mM EDTA, 1× HALT (Roche) protease inhibitors, PhosSTOP tabs). A 1.5 mg quantity of protein from each sample was added to 15 μl of washed RFP–TRAP beads (ChromoTek, number rta) and incubated for 2 h while rotating at 4 °C. Beads were washed three times with lysis buffer and eluted in 1× LDS loading dye at 94 °C for 5 min. For Flag–LC3B immunoprecipitation, 1.5 mg of protein from each sample was added to 20 μl of washed Pierce anti-Flag beads (number A36797) and incubated for 2 h while rotating at 4 °C. Beads were washed three times with lysis buffer and eluted in 1× LDS loading dye at 94 °C for 5 min.

　　A detailed protocol can be found at https://doi.org/10.17504/protocols.io.yxmvm3y8nl3p/v1. Corresponding cells were plated onto 96-well plates 1 day before the nutrient stress. The cells were washed twice with PBS and resuspended in DMEM or EBSS to start the 16-h starvation. After starvation, cells were treated with trypsin and quenched with phenol red-free DMEM. Cells were filtered and analysed by flow cytometry (Attune NxT, Thermo Fisher) using the high-throughput autosampler (CyKick). The data were processed by FlowJo software and plotted using GraphPad Prism.

　　Protocols for microscopy can be found at https://doi.org/10.17504/protocols.io.5jyl8pj9dg2w/v1. For fixed cells, cells were plated onto 18- or 22-mm glass coverslips (No. 1.5, 22 × 22-mm glass diameter, VWR 48366-227) the day before nutrient stress. DMEM was removed and cells were washed three times with DPBS, followed by resuspension in EBSS with the appropriate inhibitor(s) (SAR405, BafA1, TAK243). After starvation treatment, cells were fixed using 4% PFA followed by permeabilization with 0.5% Triton-X100. Cells were blocked in 3% BSA for 30 min, followed by incubation in primary antibodies (1:200 dilution) for 1 h at room temperature. Cells were washed three times with DPBS + 0.02% Tween-20, followed by incubation in secondary (Alexafluor conjugated 1:200 dilution) secondary antibodies for 1 h at room temperature. Coverslips were then washed three times with DPBS and 0.02% Tween-20 and mounted onto glass slides using mounting medium (Vectashield H-1000) and sealed with nail polish. The cells were imaged using a Yokogawa CSU-W1 spinning-disc confocal system on a Nikon Ti motorized microscope equipped with a Nikon Plan Apo 100×/1.40 NA objective lens, and a Hamamatsu ORCA-Fusion BT CMOS camera. For the analysis, equal gamma, brightness and contrast were applied for each image using FiJi software. For quantification, at least three separate images were quantified for the number of mNEON puncta and nuclei. For live cells, mCherry–LC3B was integrated into HEK293 cells containing an endogenous mNEON tag on YIPF4. Cells were selected with puromycin to obtain a pure population. After selection, cells were plated onto glass-bottom dishes the day before imaging. A 2 h before imaging, DMEM was removed, and cells were resuspended in EBSS to initiate autophagy. The cells were imaged using a Yokogawa CSU-W1 spinning-disc confocal system on a Nikon Ti motorized microscope equipped with a Nikon Plan Apo 100×/1.40 NA objective lens, and a Hamamatsu ORCA-Fusion BT CMOS camera, and a live-cell chamber with temperature and carbon dioxide control. For analysis, equal gamma, brightness and contrast were applied for each image using FiJi software. Quantification of the number of ATG9 puncta (objects per cell) was carried out on four or more biological replicates using Cell Profiler. Pixel size 2–15 was used to identify ATG9 vesicles, followed by normalization to cell number. Plots were created and statistical analyses were carried out using Graphpad Prism.

　　Protocols for proteomics as used here are available at https://doi.org/10.17504/protocols.io.yxmvm32nbl3p/v1 and https://doi.org/10.17504/protocols.io.dm6gp3jb1vzp/v1.

　　Cells were cultured to 70% confluency and washed with PBS three times. Cells were lysed in urea denaturing buffer (8 M urea, 150 mM NaCl, 50 mM EPPS pH 8.0, containing mammalian protease inhibitor cocktail (Sigma) and PhosSTOP) Cell lysates were collected by cell scrapers and sonicated on ice for 10 s at level 5, and the resultant extracts were clarified by centrifugation for 10 min at 15,000g at 4 °C. Lysates were quantified by the BCA assay and about 50 μg of protein was reduced with TCEP (10 mM final concentration for 30 min) and alkylated with chloroacetamide (20 mM final concentration) for 30 min. Proteins were chloroform–methanol precipitated using the SL-TMT protocol34, reconstituted in 200 mM EPPS (pH 8.5), digested by LysC for 2 h at 37 °C (1:200 wt/wt LysC/protein) and then treated with trypsin overnight at 37 °C (1:100 wt/wt trypsin/protein). about 25 μg of protein was labelled with 62.5 μg of TMT or TMTpro for 120 min at room temperature. After a labelling efficiency check, samples were quenched with hydroxylamine solution at about 0.3% final (wt in water), pooled and desalted by C18 solid-phase extraction (Sep-Pak, Waters). Pooled samples were offline fractionated with basic reverse-phase liquid chromatography (LC) into a 96-well plate and combined for a total of 24 fractions35 before desalting using a C18 StageTip (packed with Empore C18; 3M Corporation), and subsequent LC–MS/MS analysis.

　　HEK293 cells (with or without amino acid withdrawal treatment) were cultured to about 70% confluency, washed twice with chilled PBS, and collected by cell scraping in PBS. Following centrifugation at 4 °C, cell pellets were lysed in a denaturation buffer (8 M urea, 150 mM NaCl, 50 mM EPPS pH 8.0, containing mammalian protease inhibitor cocktail (Sigma), and PhosSTOP) by sonication (three times at level 5 for 5 s, with a 30 s rest on ice). Cell extracts were clarified by centrifugation for 10 min at 15,000g at 4 °C. Lysates were quantified by BCA and protein was reduced with TCEP (5 mM final concentration for 30 min), alkylated with IAA (10 mM final concentration) in the dark for 30 min, and quenched with DTT (5 mM final concentration) for 30 min. A 100 μg quantity of protein was methanol–chloroform precipitated using the SL-TMT protocol36, reconstituted in 100 mM EPPS (pH 8.5 at 1 mg ml−1), digested by LysC for 2 h at 37 °C (1:100 wt/wt LysC/protein) and then by trypsin overnight at 37 °C (1:100 wt/wt trypsin/protein). A 30 μg quantity of protein digests was acidified with formic acid to pH ≈ 3–3.5, desalted using a C18 StageTip (packed 200-μl pipette tip with Empore C18; 3M Corporation), and subjected to data-independent acquisition (DIA) LC–MS/MS analysis.

　　For APEX2 proteomics, cells expressing various APEX2–Flag fusions were processed as described previously20. To induce proximity labelling in live cells, cells were incubated with 500 μM biotin phenol (LS-3500.0250, Iris Biotech) for 1 h and treated with 1 mM H2O2 for 1 min, and the reaction was quenched with three washes of 1× PBS supplemented with 5 mM Trolox, 10 mM sodium ascorbate and 10 mM sodium azide. Cells were then collected and lysed in radioimmunoprecipitation assay (RIPA) buffer. To enrich biotinylated proteins, about 2 mg of cleared lysates was subjected to affinity purification by incubation with streptavidin-coated agarose beads (catalogue no. 88817, Pierce) for 1.5 h at room temperature. Beads were subsequently washed twice with RIPA buffer, once with 1 M KCl, once with 0.1 M NaCO3, once with PBS and once with water. For proteomics, biotinylated protein bound to the beads was reduced using TCEP (10 mM final concentration) in EPPS buffer at room temperature for 30 min. After reduction, samples were alkylated with the addition of chloracetamide (20 mM final concentration) for 20 min. Beads were washed three times with water. Proteins bound to beads were then digested with LysC (0.5 μl) in 100 ml of 0.1 M EPPS (pH 8.5) for 2 h at 37 °C, followed by trypsin overnight at 37 °C (1 μl). To quantify the relative abundance of individual protein across different samples, each digest was labelled with 62.5 mg TMT11 or TMT16pro reagents for 2 h at room temperature (Thermo Fisher), mixed, and desalted with a C18 StageTip (packed with Empore C18; 3M Corporation) before SPS-MS3 analysis on an Orbitrap Fusion Lumos Tribrid Mass Spectometer (Thermo Fisher) coupled to a Proxeon EASY-nLC 1200 LC pump (Thermo Fisher). Peptides were separated on a 100-μm-inner-diameter microcapillary column packed with about 35 cm of Accucore150 resin (2.6 μm, 150 Å, Thermo Fisher) with a gradient consisting of 5%–21% (ACN, 0.1% FA) over a total 150-min run at about 500 nl min−1 (ref. 37). The instrument parameters for each experiment are provided below.

　　Samples were analysed on an Orbitrap Fusion Lumos Tribrid Mass Spectrometer coupled to a Proxeon EASY-nLC 1200 pump (Thermo Fisher). Peptides were separated on a 35-cm column packed using a 95- to 110-min gradient. MS1 data were collected using the Orbitrap (120,000 resolution). MS2 scans were carried out in the ion trap with CID fragmentation (isolation window 0.7 Da; rapid scan; NCE 35%). Each analysis used the Multi-Notch MS3-based TMT method38, to reduce ion interference compared to MS2 quantification, combined in some instances with newly implemented Real Time Search analysis39,40, and with the FAIMS Pro Interface (using previously optimized 3 CV parameters (−40, −60, −80) for TMT multiplexed samples41). MS3 scans were collected in the Orbitrap using a resolution of 50,000, and NCE of 65 (TMT) or 45 (TMTpro). The closeout was set at two peptides per protein per fraction, so that MS3 scans were no longer collected for proteins having two peptide–spectrum matches that passed quality filters.

　　Samples were analysed on an Orbitrap Exploris 480 Mass Spectrometer coupled to a Proxeon EASY-nLC pump 1000 (Thermo Fisher). Peptides were separated on a 15-cm column packed with Accucore150 resin (150 Å, 2.6-mm C18 beads Thermo Fisher) using an 80-min acetonitrile gradient. MS1 data were collected using the Orbitrap (60,000 resolution, 350–1,050 m/z, 100% normalized AGC, maxIT set to auto). DIA MS2 scans in the Orbitrap were carried out with overlapping 24-m/z windows for the first duty cycle (390–1,014 m/z) and for the second duty cycle (402–1,026 m/z) with 28% NCE, 30,000 resolution, for fixed 145–1,450 m/z range, 1,000% normalized AGC, and a 54-ms maxIT MS1 survey scan was carried out following each DIA MS/MS duty cycle.

　　Mass spectra were converted to mzXML and monoisotopic peaks were reassigned with Monocole42 and then database searched using a Comet-based method43,44 or Sequest-HT using Proteome Discoverer (v2.3.0.420 – Thermo Fisher). Database searching included all canonical entries from the Human reference proteome database (UniProt Swiss-Prot – 2019-01; https://ftp.uniprot.org/pub/databases/uniprot/previous_major_releases/release-2019_01/) and sequences of common contaminant proteins. Searches were carried out using a 20-ppm precursor ion tolerance, and a 0.6 Da product ion tolerance for ion trap MS/MS was used. TMT tags on lysine residues and peptide N termini (+229.163 Da for Amino-TMT or +304.207 Da for TMTpro) and carbamidomethylation of cysteine residues (+57.021 Da) were set as static modifications, and oxidation of methionine residues (+15.995 Da) was set as a variable modification. Peptide–spectrum matches were filtered to a 2% false discovery rate (FDR) using linear discriminant analysis as described previously43 using the Picked FDR method45, and proteins were filtered to the target 2% FDR level. For reporter ion quantification, a 0.003-Da window around the theoretical m/z of each reporter ion was scanned, and the most intense m/z was used. Peptides were filtered to include only those peptides with >在所有TMT通道中有200个求和信号比率。在MS1隔离窗口中，隔离纯度至少为0.5（50％），用于分析的样品，而无需在线实时搜索。对于每种蛋白质，将过滤的肽 - 光谱匹配TMT或TMTPRO原始强度匹配，并将Log2归一化以创建蛋白质定量值（加权平均值）。对于蛋白质TMT定量，将TMT通道标准化为求和（蛋白质丰度实验）46或中位数（接近标记实验）47 TMT强度47 TMT强度。

　　使用MSCONVERT48转换为MSCONVERT48（仅在10-PPM质量误差下重叠）将质谱转换为MZML。使用Uniprot条目（UP000005640 [9606]）使用DIA-NN49处理MZML文件。对于DIA-NN，使用了以下参数：胰蛋白酶特异性（[RK]/P），启用N-Term蛋氨酸切除，在半胱氨酸上对氨基氨基甲基化的固定修饰，在无库模式下，基于深度学习的光谱和RTS启用，MBR启用，MBR启用，前体FDR 1％过滤器，以及使用Robust LC（较高的Precision）。使用report.pg_matrix.tsv输出DIA-NN的输出，我们根据复制强度（在至少两个生物学重复中观察到），计算了未处理和氨基酸戒断治疗条件（n = 4）的复制的平均强度，这些强度用于使用蛋白质统计型蛋白质组方法15。

　　使用学生的t检验比较了归一化的log2蛋白报道器离子强度，并使用Benjamini – Hochberg调整校正了结果P值。使用结果Q值和平均折叠变化在R中生成火山图和其他数据可视化。亚细胞列表的注释源自参考文献。14，名称源自参考。32。额外的细胞质蛋白和高尔基跨膜数注释来自Uniprot。将UNIPROT的基因本体学注释附加到MS数据中，以进行Fisher的精确测试，以识别基因本体学富集术语（通过Benjamini – Hochberg调整校正）。使用先前描述的方法15,50估算蛋白质组标尺值。未经处理的WT TMT通道对MS1前体区域（TMTWT/UT/TMTALL×MS1AREA）的比例贡献汇总到其成分肽的蛋白质水平。然后，使用所得蛋白值计算基于TMT的蛋白质组蛋白蛋白绝对丰度估计值。为了进行成像定量，使用GraphPad Prism9计算了Mann -Whitney P值。除非另有说明，否则p值<0.05被认为是显着的。使用Wilcoxon测试进行了隔室蛋白质拷贝数等级测试以计算P值。使用R（4.1.3），Rstudio IDE（2021.09.3 Build 396，Potit）和GraphPad Prism9在Adobe Illustrator中生成所有数据数字。

　　为了生成我们的盖列表，我们在自噬功能（WT）或自噬缺陷型（ATG7 - / - 或FIP200 - / - ）细胞中使用了已知的自噬通量。对于每个已知的自噬通量，使用条件中值Z评分。从这些蛋白质状况的中位数中，我们采用了已知蛋白质子集的中位数值，以估算中位数以建立共识曲线，该概况类似于“蛋白质相关分析” 51。如所预测的，使用已知自噬蛋白的共识曲线中位数值，然后我们计算了数据集中每个蛋白的RMSE。

　　通过计算每种定量蛋白质的RMSE，我们基于在数据集中的最低RMSE的前10％的蛋白质中生成了两种不同的饥饿条件下的盖子。The 10% cutoff aligns well with the rightmost tail of the density plot for the known autophagy fluxers and the top 30 autophagy factors from Fig. 2. Although the resulting ‘autophagy’ candidate list provides a defined collection of autophagy substrates, the RMSE calculation averages the error across a protein’s abundance profile, potentially enabling some proteins that vary from the consensus profile in a single condition to make the candidate列表。同样，尽管有很大程度上遵循已知的自噬磁通共识概况，但某些具有较高复制差异的自噬底物可能并不需要截止。

　　为了确定顶级候选自噬货物的优先级，我们根据蛋白质的饥饿和自噬周转率对蛋白质进行了排名（图1）和与ATG8机械的接近度（图2）。为了计算饥饿和自噬依赖性营业额的等级，我们根据较小的wt log2折叠的绝对值来确定优先级值，这是蛋白质丰度的绝对值，蛋白质丰度（从ebss/未处理的log2 [fc（ebss/ut）]≤0或atg7 - / - log2 [fc（ebss/ut）]log2 [fc（ebss/ut）]用于变化≥0（满足两个标准时）。不符合这两个标准的蛋白质被分配为0优先级。然后以降序排列优先级值，并将蛋白质缩放（蛋白质等级/实验中总蛋白质的数量）。分别计算出HELA和HEK293数据的缩放等级，并使用至少一个数据集中的最小缩放级别。根据优先级重新排序蛋白质并将排名缩放，结合两个数据集以总结图1的发现。对于ATG8接近等级，我们根据log2折叠的较小蛋白质的较小蛋白质的蛋白质丰度变化确定了优先级值，从wt ebss+bafa1/droveated for log 2 for log2 [fc+bafa in+bafa+baf aa+bafa+bafa+baf aa+baf a，ATG8 LDS突变log2 [FC（EBSS+BAFA1/UT）] - WT log2fc [（EBSS+BAFA1/UT）]用于更改≤0（仅在满足两个标准时）。如上所述，不符合两个标准的蛋白质被分配为0优先级。使用优先级值，分别计算了Apex2 – Gabrapl2和Apex2 – Map1lc3b实验的缩放等级，为此，使用了至少一个实验中的最小缩放级别。基于优先级重新排序蛋白质，并将排名缩放，结合了两个数据集，以总结图2的发现。优先考虑表现出自噬和饥饿依赖性的周转率和与ATG8相关联的候选者的优先级。 我们将图1和图2的缩放等级求和，以生成一个总的等级值，我们通过上升顺序排序以生成我们最终排名的候选列表。要成为最终排名列表中的候选者，必须在图1实验（HELA或HEK293）中至少在一个实验中鉴定该蛋白质，并从图2实验（Apex2 -Gabarapl2和Apex2 – Map1lc3b）中鉴定出一个实验。从ILIR自噬数据库（http://repeat.biol.ucy.ac.ac.cy/ilir/)21匹配LIR图案。已知的自噬蛋白来自参考。10。尽管RMSE方法可能无法捕获所有自噬底物，但是优先考虑的帽子集合使我们能够在养分胁迫下定义大噬菌的选择性。

　　有关研究设计的更多信息可在与本文有关的自然投资组合报告摘要中获得。