1. ###### tags: `Sprint Diary` # 4DN2 SPRINT DIARY - Q1S3 ## 14 Feb. 2023 / 11:30 AM to 28 Feb. 2023 ## Timeline [Airtable Timeline](https://airtable.com/shrLOuMSWU0UlAA9F) ### Attendees Yu, Emily, Trevor, Johan, Jiangyuan, Sasha, Vedat, Nezar, ## Resources * Sprint Slides for Figures: https://docs.google.com/presentation/d/18Jo4SmvIzYtHb1dFxO4IUFcQlSIOIzu2-QbkI2RTLoE/edit?usp=sharing * Kanban Board: https://airtable.com/shrNznOkqe3JGugKK * Github: https://github.com/abdenlab/4dn2-sprints.git * Whiteboard: https://miro.com/app/board/uXjVPoAbEpc=/?share_link_id=737466028503 * Action Items: ## Sprint Agenda **Previous Items:** ### Planned: * Trajectories of RNAseq; Relationship to epigenetic compartmentalization. * Insulation Scores and Short range interactions, TADs, Hi-C Low dimensional features * ATAC-Seq peaks across the developmental stages. * In Progress: * Consolidated Epigenetic tracks * Clustering discussion, focus on joint clusters, GSEA on joint clusters. if polycomb is established, expression will go down. * Improvements on IPG Clustering * Dot calling for newly mapped Hi-C Lanes: * Standard dot calling on the megamap! might give a good set of union_dots. * RNAseq: Pairwise DEx * Number of nearest neighbors, UMAP * avg exp levels in different cond. may be more variable, something that is orthagonal to the timeseries. * Coloring by the IPGs on UMAPs. ## Huddle Day Agenda Feb 14th, 2023 * 11.30am - 12pm Morning Huddle * 12pm - 1pm 4DN2-CSG Meeting (JC + Q1S2 updates) * Afternoon for breakouts. ## Morning Huddle: * Discussion notes: * Dot calling, parameters for singletons! * Yu might improve PCA by adding third dimension. * Yu will try other kinds of annotations, such as GSEA. * From JC paper, It's worth doing a classic A-B transitions across stages and mapping DEGs in A-B * 50kb gene transcript level, open chromatin level, At least with 2 conditions, making a union list of ATAC peaks. A refence for accesible loci. * Epigenetics data to characterize accesible regions, Despite not having those at right stages, we can measure the proeprties of atac peaks in 5 stages. Prelim analysis with first two stages. ## CSG meeting Emily presenting Alavattam et al, 2023. [Notes](https://docs.google.com/document/d/1J1DfjImyK3x6wqbdgDAdv7OkIfZL7IeheUO3VF_pPIo) ## Afternoon Huddle 02/14/2023 ### With Yu high rez IPG, with Cis-Trans. done by Sasha. She can use data from there to incorporate into RNA trajectories. We havent generated many standard things, insulation scores, classic eigen vectors. expected SCC?, If we get final coolers from Johan. Rep1 all merged, all five stages. We can use cooltools to generate those. Based on TSS, incorporate Hi-C and RNAseq. With insulation scores you can annotate genes with that??. Include PCA with mHEP, PC3 1- GREAT IDEA!!! Cell cycle score, markers for each cell cycle. To see the proportion of the each cell cycle, we cAN DO do for each stage of cell cycle? ?relative expression levels of S vs G1 vs G2 - great IDEA!!! 2- mHEP in sc-RNA seq to compare bulk to SC RNAseq. 3- 3rd dimension in PCAs. 4- GSEA on RNAseq, on IPG clusters. What does GSEA CLI use? Provide Links for CLI to use. ### With Jiangyuan Clusters 2x2, each 5kb pixel. Venn diagram problem, if you set a certain treshold, it's everywhere. Compare union list from megamap to current approach. Relative loop strentgh at 5 conditions. Goals to assign clusters to reference points. Union list to track loop strenth. In a single loop dependent manner. Soft assignments is a way of doing it instead of a pile up. ITS vs Delta E1 If you choose a cut-off to ignore, you can look at correlation fo coefficients, whether if it improves when we ignore the housekeeping genes that do not change. ITS, determine theta, calculate R^2, compare R^2's for all 16 of those. 4x4 plot of theta vs R^2. How strong the correlation is!!! Do first! UMAP clusters: bigger and smaller compartmets changing compartments/clusters. most on Chr 9. Very interesting location. Q arm? Can you recalculate eigen vectors on Hi-C 3.0? When we have latest data, we produce latest set of EI vectors. Don't use 100kb for Cis! Cis interactions require higher resolution. Trans signal is shallow, hard to fit into memory. Sasha calc trans at 50kb cis on 5-10kb, project on Trans eigen coordinate system, then you have nicely aligned Cis eigen vectors. Something might be wrong with Chr9 eigen vectors. intact hi-c if they can be public! ## Sprint Notes: * Finalized Hi-C maps! One final frozen version of Hi-C, and generate everything we generate, compartments Eig. at Cis and Trans, insulation scores at diff resolution, dot calls. * While filtering HI-C maps, all the same cell types, no structural variations, a common mask could be used for 5 stages, less stringent, helps with analysis. * using synthetic data see whether Hi-C maps look similar to those that divide. ## Huddle Day Agenda Feb 21st, 2023 * 11.30am - 12pm Morning Huddle Yu, Nezar, Trevor, Emily, Jiangyuan, Sasha. Mohan is repeating the diff protocol. Yu: Present RNAseq analysis work for CSG meeting. Summary of RNAseq analysis. GSEA comparison on diff vs pluripotency, groupings of gene sets, if genome structure is more important in one type? Emily: Talked Nezar and Mohan abt chip-seq. Follow the format. Biggest challenge: communication, solved. Trevor: Finishing up pieces to release HG python, feeling demotv. on release, want to do visualization instead of HG python. Sasha: Projections cis trans -25kb- 50kb, SAsha doing 50kb projections, Eigen vectors are projected! Exploring from last week, how to do NonNegMatrixF, what Epigenetic signal to use? 3 tracks. Emily has track databases. Jiangyuan: Merged the dots from individual stages to megamaps! Those megamap specific dots are close to diagonal. ## Standup Notes from Previous Sprint | Member | Date | What are you working on this sprint? | What progress have you had so far? | What will you do next? | Anything blocking your progress? | | -------------------- | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------- | | Nezar Abdennur | 2023-02-21 | • Klein et al. data analysis for ROCCS group<br>• Experimental design for center | Had a meeting with Mohan and Emily last week to prioritize ChIP/C&amp;R targets based on budget constraints and new differentiation batch to be started. I created an <#C04QQF71S6L\|experiments> channel to continue the discussion. We decided to include the original essential marks (link to spreadsheet is bookmarked in the channel), and include H3K36me3 on Emily's suggestion, as well as H3K9me2 now that we seem to have found a validated antibody. | Finish up the ROCCS data processing and Resgen views and draft an email to the group summarizing the findings. | Just time. Nothing else. | | Nezar Abdennur | 2023-02-23 | • Klein et al. data analysis for ROCCS group<br>• Experimental design for center | Sent a summary of the results to the ROCCS group this morning. Views are here: <https://resgen.io/4dn2csg/ROCCS/views/> | A couple more ChIP datasets to map and upload. Read Dave's replies and respond. | Not at the moment. | | Emily Navarrete | 2023-02-23 | Rewriting code for histone mark pileups | I think I’ve gotten the pileups saved properly to hdf5 files, now I need to check that is true and rewrite my plotting code | now I need to check that is true and rewrite my plotting code | balancing projects | | Trevor Manz | 2023-02-23 | Finishing the migration of `hg` to `higlass/higlass-python` | • merged initial transition <https://github.com/higlass/higlass-python/pull/96><br>• migrated build system to hatch | Finish migrating the documentation (current progress is available at <http://higlass.github.io/higlass-python\|higlass.github.io/higlass-python>) and make an initial release | Not at the moment. | | YU FU | 2023-02-24 | • Make converted mcool public access for browsing and link them to google sheet<br>• Working on Gene set enrichment analysis (wanna do network graph) for easy browsing interesting genes in each expression group<br>• Putting together my codes and slides for submission and next week meeting | • Done journal club (very good data)<br>• mcool transfer in process<br>• Made volcano plots for DEGs and start gene set enrichment for DEGs | • continue on gene set enrichment analysis <br>• link mcool to google sheet<br>• prep for presentation | No | | Aleksandra Galitsyna | 2023-02-24 | IPGs for updated datasets (14 Feb rom Johan; Rep1 Hi-C 3.0 completed and Rep2 draft) | EVs and projections on mirnymachines (important for <#C04E01PGXUH\|analysis-vis> and <#C02G03WQHCY\|datasets>)<br>• Location of the updated config file with the new data (UPD ~14 Feb 2023 by Johan &amp; Sasha) : `/net/levsha/share/lab/4dn2csg/hepatic_diff_2022/ipg/hic_hic3.0/config_7+.yml`<br>• trans EVs: `/net/levsha/share/lab/4dn2csg/hepatic_diff_2022/ipg/hic_hic3.0/inspectro/results`<br>• 50-50 Kb projections from cis to trans: `/net/levsha/share/lab/4dn2csg/hepatic_diff_2022/ipg/hic_hic3.0/projections_cis-trans/results`<br>• 25-50 Kb projections<br>• 10-50 Kb projections<br>• 5-50 Kb projections | Clustering of IPGs and QC | - |