AHA BCVS GTEx
Brian Gural
7/3/2023
Project Goals
Estimate cellular composition in human hearts from bulk RNAseq
# bulk
bulk <- read.csv("gtex/data/processed/internal/gtex_lv_counts_summed.csv", check.names = F, row.names = 2)
bulk <- bulk[,-1]
# Single cell of all datasets combined
sn <- LoadH5Seurat("gtex/data/processed/internal/sn_gtex_no_de.h5seurat")
print(sn)## An object of class Seurat
## 12839 features across 36574 samples within 1 assay
## Active assay: RNA (12839 features, 0 variable features)
These plots show how distinct and segregated some of our datasets are. It’s important to note that Wu and Martini both are specifically enriched for immune cells. And Tabula Muris is deficient in cardiomyoctes generally. These clusters aren’t yet annotated, but keep that in mind when we assign cell-types to these clusters later. I’m still concerned about the limited overlap in cells in clusters 0 and 2. I’d like to find more sc/sn datasets that overlap between the clusters so we can better integrate the data.
I’d also like to note that I’ve run the pipeline in two additional ways to check things. I tried each combination of datasets (8 total if Rau is included every time) and also downsampled the counts to 5k (to check if the difference in the “Counts by origin” plot effect things). For the former, more datasets improve the results (I’ll include a plot from this analysis in an email) and for the later, there isn’t any notable difference when using more uniform counts between the datasets.
## Gene Ensembl.ID Gene.type Leiden.cluster Cell.type N.nuclei
## 1 FGF12 ENSG00000114279 protein_coding 0 Cardiomyocyte I 31408
## 2 MLIP ENSG00000146147 protein_coding 0 Cardiomyocyte I 31408
## 3 RBM20 ENSG00000203867 protein_coding 0 Cardiomyocyte I 31408
## 4 FHL2 ENSG00000115641 protein_coding 0 Cardiomyocyte I 31408
## 5 TECRL ENSG00000205678 protein_coding 0 Cardiomyocyte I 31408
## 6 CDH2 ENSG00000170558 protein_coding 0 Cardiomyocyte I 31408
## Pct...0..target Pct...0..non.target Avg.Expr...target Avg.Expr...non.target
## 1 0.976 0.010 3.91 0.03
## 2 0.953 0.005 3.21 0.02
## 3 0.936 0.007 2.81 0.02
## 4 0.959 0.013 3.50 0.03
## 5 0.893 0.003 2.49 0.01
## 6 0.888 0.004 2.46 0.01
## limma.voom..log.FC. limma.voom..P.Value limma.voom..Adjusted.P.Value AUC
## 1 8.65 2.10e-39 1.21e-37 0.99
## 2 8.35 3.77e-44 3.54e-42 0.97
## 3 8.22 1.41e-44 1.39e-42 0.96
## 4 8.03 1.15e-32 3.09e-31 0.98
## 5 7.99 2.43e-51 4.30e-49 0.94
## 6 7.90 2.77e-50 4.41e-48 0.94As we talked about last time, AUCell assigns an AUC score (think enrichment) for each cell based on expression and cell types with clear bimodal distributions are retained. The second part, retaining only cell-types with distinct groups, is a change from the last workflow that has been very helpful. Here’s how those AUC distributions look for the cell types we retained:
We set a threshold (0.4 in this case) and assign cell types based on that. Here’s how our final labels look, along with some canonical markers to confirm our annotations:
The ambient expression of markers in clusters dominated by the Rau dataset make me think that there is still ambient RNA (RNA from cells that broke open during handling and are present in most droplets). I’ll just need to go back to the raw dataset, remap it, and apply some newer methods. It would just take a bit of time and I didn’t want to delay providing an update.
Deconvolution
Composition estimates on pure cell fractions
Experimentally purified, bulk sequenced cell types have been used as a ground truth. This is how I’ve been testing accuracy of the methods. Specifically, I’ve been trying to minimize the Aitchison distance, a measure of composition distance that is apparently widely used in fields like geology, ecology, and microbiome research. The plot I mentioned about dataset combinations uses that as a metric.
The fractions are from Christoph and a publicly available dataset, which is why you can see two distinct groups in the endothelial cell graph.
## Creating Relative Abudance Matrix...## Creating Variance Matrix...## Creating Library Size Matrix...## Used 12821 common genes...## Used 4 cell types in deconvolution...## GTEX-111FC has common genes 12686 ...## GTEX-111YS has common genes 12566 ...## GTEX-1122O has common genes 12625 ...## GTEX-117YW has common genes 12542 ...## GTEX-117YX has common genes 12513 ...## GTEX-11DXX has common genes 12579 ...## GTEX-11DXY has common genes 12544 ...## GTEX-11DXZ has common genes 12630 ...## GTEX-11EM3 has common genes 12564 ...## GTEX-11EMC has common genes 12689 ...## GTEX-11GS4 has common genes 12664 ...## GTEX-11GSP has common genes 12600 ...## GTEX-11I78 has common genes 12651 ...## GTEX-11LCK has common genes 12641 ...## GTEX-11O72 has common genes 12650 ...## GTEX-11ONC has common genes 12524 ...## GTEX-11TT1 has common genes 12505 ...## GTEX-11TUW has common genes 12687 ...## GTEX-11ZUS has common genes 12597 ...## GTEX-11ZVC has common genes 12622 ...## GTEX-1211K has common genes 12615 ...## GTEX-1212Z has common genes 12622 ...## GTEX-12584 has common genes 12603 ...## GTEX-12696 has common genes 12641 ...## GTEX-1269C has common genes 12580 ...## GTEX-12BJ1 has common genes 12568 ...## GTEX-12WSC has common genes 12621 ...## GTEX-12WSD has common genes 12662 ...## GTEX-12WSG has common genes 12676 ...## GTEX-12WSK has common genes 12641 ...## GTEX-12WSN has common genes 12584 ...## GTEX-12ZZW has common genes 12606 ...## GTEX-12ZZY has common genes 12574 ...## GTEX-12ZZZ has common genes 12637 ...## GTEX-13112 has common genes 12661 ...## GTEX-1313W has common genes 12706 ...## GTEX-1314G has common genes 12650 ...## GTEX-131XE has common genes 12637 ...## GTEX-131XF has common genes 12584 ...## GTEX-131XG has common genes 12621 ...## GTEX-131XH has common genes 12484 ...## GTEX-132AR has common genes 12634 ...## GTEX-132NY has common genes 12649 ...## GTEX-132QS has common genes 12629 ...## GTEX-1339X has common genes 12573 ...## GTEX-1399R has common genes 12567 ...## GTEX-1399S has common genes 12564 ...## GTEX-1399T has common genes 12394 ...## GTEX-1399U has common genes 12586 ...## GTEX-139YR has common genes 12548 ...## GTEX-13CF2 has common genes 12579 ...## GTEX-13CF3 has common genes 12644 ...## GTEX-13D11 has common genes 12622 ...## GTEX-13FH7 has common genes 12655 ...## GTEX-13FHP has common genes 12389 ...## GTEX-13FLV has common genes 12643 ...## GTEX-13FTW has common genes 12547 ...## GTEX-13FTZ has common genes 12606 ...## GTEX-13JVG has common genes 12490 ...## GTEX-13N11 has common genes 12643 ...## GTEX-13N1W has common genes 12562 ...## GTEX-13NYB has common genes 12328 ...## GTEX-13NZB has common genes 12627 ...## GTEX-13O1R has common genes 12626 ...## GTEX-13O21 has common genes 12506 ...## GTEX-13O3O has common genes 12469 ...## GTEX-13O3P has common genes 12561 ...## GTEX-13O61 has common genes 12659 ...## GTEX-13OVG has common genes 12633 ...## GTEX-13OVL has common genes 12637 ...## GTEX-13OW5 has common genes 12482 ...## GTEX-13OW6 has common genes 12598 ...## GTEX-13OW8 has common genes 12560 ...## GTEX-13PVQ has common genes 12512 ...## GTEX-13PVR has common genes 12490 ...## GTEX-13QBU has common genes 12631 ...## GTEX-13RTJ has common genes 12442 ...## GTEX-13S86 has common genes 12498 ...## GTEX-13SLX has common genes 12661 ...## GTEX-13U4I has common genes 12641 ...## GTEX-13VXT has common genes 12625 ...## GTEX-13W3W has common genes 12675 ...## GTEX-13X6H has common genes 12523 ...## GTEX-13X6I has common genes 12654 ...## GTEX-13X6K has common genes 12692 ...## GTEX-13YAN has common genes 12460 ...## GTEX-145LV has common genes 12528 ...## GTEX-145MN has common genes 12622 ...## GTEX-145MO has common genes 12571 ...## GTEX-146FH has common genes 12636 ...## GTEX-146FR has common genes 12613 ...## GTEX-147F4 has common genes 12517 ...## GTEX-147JS has common genes 12508 ...## GTEX-148VI has common genes 12486 ...## GTEX-148VJ has common genes 12394 ...## GTEX-1497J has common genes 12584 ...## GTEX-14A5I has common genes 12264 ...## GTEX-14A6H has common genes 12655 ...## GTEX-14BIL has common genes 12302 ...## GTEX-14BIN has common genes 12214 ...## GTEX-14C38 has common genes 12576 ...## GTEX-14C39 has common genes 12480 ...## GTEX-14C5O has common genes 12324 ...## GTEX-14DAQ has common genes 12680 ...## GTEX-14E1K has common genes 12647 ...## GTEX-14E6C has common genes 12580 ...## GTEX-14E6D has common genes 12393 ...## GTEX-14E7W has common genes 12429 ...## GTEX-14H4A has common genes 12499 ...## GTEX-14JG1 has common genes 12587 ...## GTEX-14JIY has common genes 12404 ...## GTEX-14PHX has common genes 12570 ...## GTEX-14PHY has common genes 12583 ...## GTEX-14PJ2 has common genes 12556 ...## GTEX-14PJ4 has common genes 12651 ...## GTEX-14PJ6 has common genes 12546 ...## GTEX-14PJM has common genes 12764 ...## GTEX-14PK6 has common genes 12549 ...## GTEX-14PN3 has common genes 12489 ...## GTEX-14XAO has common genes 12581 ...## GTEX-15CHC has common genes 12633 ...## GTEX-15CHQ has common genes 12547 ...## GTEX-15CHR has common genes 12579 ...## GTEX-15DDE has common genes 12628 ...## GTEX-15DYW has common genes 12522 ...## GTEX-15EOM has common genes 12607 ...## GTEX-15ER7 has common genes 12688 ...## GTEX-15RIE has common genes 12609 ...## GTEX-15RJ7 has common genes 12654 ...## GTEX-15RJE has common genes 12697 ...## GTEX-15SHU has common genes 12481 ...## GTEX-15SHW has common genes 12353 ...## GTEX-16BQI has common genes 12577 ...## GTEX-16MT8 has common genes 12472 ...## GTEX-16NPX has common genes 12666 ...## GTEX-16XZY has common genes 12564 ...## GTEX-16YQH has common genes 12620 ...## GTEX-17F9E has common genes 12472 ...## GTEX-17F9Y has common genes 12388 ...## GTEX-17GQL has common genes 12639 ...## GTEX-17HGU has common genes 12626 ...## GTEX-17HHY has common genes 12550 ...## GTEX-17HII has common genes 12623 ...## GTEX-17JCI has common genes 12513 ...## GTEX-17KNJ has common genes 12668 ...## GTEX-17MF6 has common genes 12643 ...## GTEX-18465 has common genes 12449 ...## GTEX-18A66 has common genes 12694 ...## GTEX-18A67 has common genes 12631 ...## GTEX-18A6Q has common genes 12663 ...## GTEX-18D9B has common genes 12576 ...## GTEX-18D9U has common genes 12481 ...## GTEX-1A3MW has common genes 12713 ...## GTEX-1A3MX has common genes 12612 ...## GTEX-1A8G6 has common genes 12409 ...## GTEX-1AX8Z has common genes 12695 ...## GTEX-1AX9I has common genes 12547 ...## GTEX-1AX9J has common genes 12499 ...## GTEX-1AX9K has common genes 12631 ...## GTEX-1AYCT has common genes 12658 ...## GTEX-1B8KE has common genes 12537 ...## GTEX-1B8L1 has common genes 12641 ...## GTEX-1B8SG has common genes 12549 ...## GTEX-1B932 has common genes 12626 ...## GTEX-1B97J has common genes 12571 ...## GTEX-1B996 has common genes 12620 ...## GTEX-1BAJH has common genes 12669 ...## GTEX-1C2JI has common genes 12521 ...## GTEX-1C64O has common genes 12584 ...## GTEX-1C6VQ has common genes 12373 ...## GTEX-1CAMS has common genes 12622 ...## GTEX-1CB4E has common genes 12664 ...## GTEX-1CB4G has common genes 12686 ...## GTEX-1CB4I has common genes 12663 ...## GTEX-1CB4J has common genes 12643 ...## GTEX-1E1VI has common genes 12596 ...## GTEX-1EH9U has common genes 12432 ...## GTEX-1EKGG has common genes 12396 ...## GTEX-1EMGI has common genes 12585 ...## GTEX-1EN7A has common genes 12630 ...## GTEX-1EU9M has common genes 12625 ...## GTEX-1EWIQ has common genes 12483 ...## GTEX-1EX96 has common genes 12468 ...## GTEX-1F48J has common genes 12646 ...## GTEX-1F5PK has common genes 12586 ...## GTEX-1F6I4 has common genes 12644 ...## GTEX-1F6IF has common genes 12688 ...## GTEX-1F75I has common genes 12579 ...## GTEX-1GF9U has common genes 12671 ...## GTEX-1GF9V has common genes 12586 ...## GTEX-1GL5R has common genes 12573 ...## GTEX-1GMR2 has common genes 12549 ...## GTEX-1GMR3 has common genes 12585 ...## GTEX-1GMRU has common genes 12701 ...## GTEX-1GN1V has common genes 12633 ...## GTEX-1GN1W has common genes 12650 ...## GTEX-1GN2E has common genes 12695 ...## GTEX-1GN73 has common genes 12634 ...## GTEX-1GTWX has common genes 12726 ...## GTEX-1GZ4I has common genes 12657 ...## GTEX-1H11D has common genes 12631 ...## GTEX-1H1CY has common genes 12475 ...## GTEX-1H1ZS has common genes 12578 ...## GTEX-1H23P has common genes 12561 ...## GTEX-1H3NZ has common genes 12551 ...## GTEX-1HBPH has common genes 12658 ...## GTEX-1HBPM has common genes 12568 ...## GTEX-1HC8U has common genes 12668 ...## GTEX-1HCU6 has common genes 12489 ...## GTEX-1HCU7 has common genes 12627 ...## GTEX-1HCUA has common genes 12661 ...## GTEX-1HCVE has common genes 12444 ...## GTEX-1HFI7 has common genes 12633 ...## GTEX-1HSGN has common genes 12609 ...## GTEX-1HSMO has common genes 12678 ...## GTEX-1HSMQ has common genes 12661 ...## GTEX-1HT8W has common genes 12626 ...## GTEX-1I1GP has common genes 12310 ...## GTEX-1I1GQ has common genes 12663 ...## GTEX-1I1GV has common genes 12587 ...## GTEX-1I1HK has common genes 12686 ...## GTEX-1I6K7 has common genes 12543 ...## GTEX-1ICG6 has common genes 12612 ...## GTEX-1ICLY has common genes 12502 ...## GTEX-1ICLZ has common genes 12613 ...## GTEX-1IDJE has common genes 12654 ...## GTEX-1IDJF has common genes 12653 ...## GTEX-1IDJH has common genes 12350 ...## GTEX-1IDJI has common genes 12609 ...## GTEX-1IDJU has common genes 12709 ...## GTEX-1IKJJ has common genes 12325 ...## GTEX-1IKK5 has common genes 12674 ...## GTEX-1IKOE has common genes 12560 ...## GTEX-1IL2U has common genes 12604 ...## GTEX-1J8EW has common genes 12659 ...## GTEX-1J8Q2 has common genes 12649 ...## GTEX-1J8Q3 has common genes 12658 ...## GTEX-1J8QM has common genes 12692 ...## GTEX-1JJ6O has common genes 12643 ...## GTEX-1JJE9 has common genes 12512 ...## GTEX-1JJEA has common genes 12603 ...## GTEX-1JK1U has common genes 12593 ...## GTEX-1JMLX has common genes 12569 ...## GTEX-1JMOU has common genes 12517 ...## GTEX-1JMPZ has common genes 12654 ...## GTEX-1JMQK has common genes 12653 ...## GTEX-1JN1M has common genes 12669 ...## GTEX-1JN6P has common genes 12273 ...## GTEX-1K9T9 has common genes 12642 ...## GTEX-1KANC has common genes 12590 ...## GTEX-1KD4Q has common genes 12372 ...## GTEX-1KD5A has common genes 12609 ...## GTEX-1KXAM has common genes 12640 ...## GTEX-1L5NE has common genes 12579 ...## GTEX-1LB8K has common genes 12559 ...## GTEX-1LC47 has common genes 12618 ...## GTEX-1LG7Y has common genes 12522 ...## GTEX-1LGOU has common genes 12579 ...## GTEX-1LGRB has common genes 12718 ...## GTEX-1LH75 has common genes 12598 ...## GTEX-1LSNM has common genes 12593 ...## GTEX-1LVA9 has common genes 12461 ...## GTEX-1LVAN has common genes 12577 ...## GTEX-1M5QR has common genes 12625 ...## GTEX-1MCC2 has common genes 12683 ...## GTEX-1MJIX has common genes 12636 ...## GTEX-1MUQO has common genes 12595 ...## GTEX-1N2EF has common genes 12475 ...## GTEX-1NV8Z has common genes 12431 ...## GTEX-1O97I has common genes 12385 ...## GTEX-1O9I2 has common genes 12582 ...## GTEX-1OFPY has common genes 12556 ...## GTEX-1OKEX has common genes 12555 ...## GTEX-1POEN has common genes 12647 ...## GTEX-1PPGY has common genes 12602 ...## GTEX-1PPH7 has common genes 12611 ...## GTEX-1PWST has common genes 12559 ...## GTEX-1QCLZ has common genes 12587 ...## GTEX-1QP29 has common genes 12549 ...## GTEX-1QP2A has common genes 12670 ...## GTEX-1QPFJ has common genes 12650 ...## GTEX-1R7EV has common genes 12606 ...## GTEX-1R9JW has common genes 12662 ...## GTEX-1R9K4 has common genes 12612 ...## GTEX-1R9K5 has common genes 12643 ...## GTEX-1R9PO has common genes 12547 ...## GTEX-1RAZS has common genes 12417 ...## GTEX-1RB15 has common genes 12591 ...## GTEX-1RMOY has common genes 12587 ...## GTEX-1RQEC has common genes 12582 ...## GTEX-1S3DN has common genes 12534 ...## GTEX-N7MS has common genes 12720 ...## GTEX-N7MT has common genes 12530 ...## GTEX-NFK9 has common genes 12668 ...## GTEX-NPJ8 has common genes 12666 ...## GTEX-O5YT has common genes 12578 ...## GTEX-O5YV has common genes 12683 ...## GTEX-O5YW has common genes 12600 ...## GTEX-OHPK has common genes 12669 ...## GTEX-OHPL has common genes 12663 ...## GTEX-OHPM has common genes 12624 ...## GTEX-OIZF has common genes 12675 ...## GTEX-OIZG has common genes 12651 ...## GTEX-OIZH has common genes 12619 ...## GTEX-OIZI has common genes 12660 ...## GTEX-OOBJ has common genes 12671 ...## GTEX-OXRK has common genes 12667 ...## GTEX-OXRL has common genes 12603 ...## GTEX-OXRO has common genes 12385 ...## GTEX-P44G has common genes 12583 ...## GTEX-P44H has common genes 12627 ...## GTEX-P4PP has common genes 12618 ...## GTEX-P4PQ has common genes 12645 ...## GTEX-P4QS has common genes 12663 ...## GTEX-P78B has common genes 12661 ...## GTEX-PLZ5 has common genes 12683 ...## GTEX-POMQ has common genes 12691 ...## GTEX-PSDG has common genes 12680 ...## GTEX-PVOW has common genes 12626 ...## GTEX-PWCY has common genes 12673 ...## GTEX-PX3G has common genes 12642 ...## GTEX-Q2AH has common genes 12610 ...## GTEX-QDVJ has common genes 12691 ...## GTEX-QDVN has common genes 12607 ...## GTEX-QEG4 has common genes 12647 ...## GTEX-QEG5 has common genes 12638 ...## GTEX-QEL4 has common genes 12604 ...## GTEX-QESD has common genes 12590 ...## GTEX-QLQ7 has common genes 12672 ...## GTEX-QMRM has common genes 12661 ...## GTEX-QV44 has common genes 12597 ...## GTEX-QVJO has common genes 12273 ...## GTEX-R45C has common genes 12573 ...## GTEX-R53T has common genes 12677 ...## GTEX-R55C has common genes 12457 ...## GTEX-R55E has common genes 12653 ...## GTEX-R55G has common genes 12688 ...## GTEX-REY6 has common genes 12591 ...## GTEX-RN64 has common genes 12590 ...## GTEX-RNOR has common genes 12529 ...## GTEX-RTLS has common genes 12580 ...## GTEX-RU72 has common genes 12646 ...## GTEX-RUSQ has common genes 12467 ...## GTEX-RWS6 has common genes 12602 ...## GTEX-RWSA has common genes 12507 ...## GTEX-S32W has common genes 12652 ...## GTEX-S33H has common genes 12573 ...## GTEX-SE5C has common genes 12646 ...## GTEX-SIU8 has common genes 12622 ...## GTEX-SNMC has common genes 12584 ...## GTEX-SUCS has common genes 12595 ...## GTEX-T2IS has common genes 12573 ...## GTEX-T6MN has common genes 12684 ...## GTEX-U3ZH has common genes 12706 ...## GTEX-U3ZN has common genes 12744 ...## GTEX-U4B1 has common genes 12607 ...## GTEX-U8XE has common genes 12551 ...## GTEX-UJHI has common genes 12571 ...## GTEX-UJMC has common genes 12682 ...## GTEX-UPK5 has common genes 12641 ...## GTEX-V1D1 has common genes 12605 ...## GTEX-V955 has common genes 12635 ...## GTEX-WEY5 has common genes 12495 ...## GTEX-WFG7 has common genes 12534 ...## GTEX-WFG8 has common genes 12580 ...## GTEX-WFON has common genes 12544 ...## GTEX-WH7G has common genes 12671 ...## GTEX-WHPG has common genes 12683 ...## GTEX-WHSE has common genes 12537 ...## GTEX-WHWD has common genes 12651 ...## GTEX-WI4N has common genes 12645 ...## GTEX-WK11 has common genes 12511 ...## GTEX-WL46 has common genes 12681 ...## GTEX-WQUQ has common genes 12697 ...## GTEX-WRHU has common genes 12556 ...## GTEX-WWYW has common genes 12672 ...## GTEX-WY7C has common genes 12676 ...## GTEX-WYJK has common genes 12684 ...## GTEX-WZTO has common genes 12712 ...## GTEX-X15G has common genes 12554 ...## GTEX-X261 has common genes 12399 ...## GTEX-X3Y1 has common genes 12659 ...## GTEX-X8HC has common genes 12518 ...## GTEX-XBEC has common genes 12567 ...## GTEX-XBED has common genes 12558 ...## GTEX-XGQ4 has common genes 12540 ...## GTEX-XPT6 has common genes 12698 ...## GTEX-XPVG has common genes 12690 ...## GTEX-XQ3S has common genes 12553 ...## GTEX-XQ8I has common genes 12399 ...## GTEX-XV7Q has common genes 12652 ...## GTEX-XXEK has common genes 12500 ...## GTEX-Y114 has common genes 12650 ...## GTEX-Y3I4 has common genes 12525 ...## GTEX-Y3IK has common genes 12576 ...## GTEX-Y5V5 has common genes 12656 ...## GTEX-Y5V6 has common genes 12656 ...## GTEX-Y8E4 has common genes 12648 ...## GTEX-YB5K has common genes 12595 ...## GTEX-YEC3 has common genes 12660 ...## GTEX-YEC4 has common genes 12681 ...## GTEX-YF7O has common genes 12660 ...## GTEX-YFC4 has common genes 12661 ...## GTEX-YJ8O has common genes 12587 ...## GTEX-ZAB4 has common genes 12662 ...## GTEX-ZAB5 has common genes 12670 ...## GTEX-ZAJG has common genes 12614 ...## GTEX-ZDTS has common genes 12628 ...## GTEX-ZDTT has common genes 12656 ...## GTEX-ZDXO has common genes 12618 ...## GTEX-ZDYS has common genes 12659 ...## GTEX-ZEX8 has common genes 12635 ...## GTEX-ZF29 has common genes 12623 ...## GTEX-ZF2S has common genes 12659 ...## GTEX-ZF3C has common genes 12472 ...## GTEX-ZG7Y has common genes 12608 ...## GTEX-ZGAY has common genes 12647 ...## GTEX-ZLFU has common genes 12600 ...## GTEX-ZPCL has common genes 12667 ...## GTEX-ZPIC has common genes 12612 ...## GTEX-ZPU1 has common genes 12649 ...## GTEX-ZQUD has common genes 12430 ...## GTEX-ZTPG has common genes 12631 ...## GTEX-ZUA1 has common genes 12584 ...## GTEX-ZV7C has common genes 12609 ...## GTEX-ZVT2 has common genes 12666 ...## GTEX-ZVZP has common genes 12560 ...## GTEX-ZYFC has common genes 12620 ...## GTEX-ZYFG has common genes 12619 ...## GTEX-ZYT6 has common genes 12554 ...## GTEX-ZYW4 has common genes 12682 ...## GTEX-ZZPU has common genes 12652 ...Composition estimates in Jensen data
Here’s the actual results you’re intereted in. The KO vs. WT labels were based on the ‘id’ column of the file you gave me (labels include WT5, AKO5, Ako_Lx1, wt_Lx1, and Ako_Lx4), rather than the ‘condition’ column (labels include: WT.sham, KO.sham, WT.MI, KO.MI, and WT.outlier). Just let me know if I got this right, since I know you had mentioned that they were switched.
##
## 0 Organ Donor (OPO) Postmortem Surgical
## 3 212 238 12##
## 0 1 2
## 3 308 154
Using DirichletReg to model compositional changes
library("DirichletReg")## Loading required package: Formula# prep DirichletReg matrix
dir.merged <- merge(decon.melt, pheno.basic, by.x = "Sub", by.y = "SUBJID")
dir.mat <- dir.merged |>
dcast(Sub + COHORT + SEX + AGE + BMI ~ CellType, value.var = "Prop")
# Convert data to DirichletRegData object
dir.mat$CellTypes <- DR_data(dir.mat[,c(6:length(dir.mat))])
dir.mat$Sub <- as.factor(dir.mat$Sub)
dir.mat$COHORT <- as.factor(dir.mat$COHORT)
dir.mat$SEX <- as.factor(dir.mat$SEX)
dir.mat$AGE <- as.numeric(dir.mat$AGE)
dir.mat$BMI <- as.numeric(dir.mat$BMI)
# Run Dirichlet regression
model.1 <- DirichReg(CellTypes ~ COHORT + SEX + AGE + BMI, data = dir.mat)
model.2 <- DirichReg(CellTypes ~ SEX + AGE + BMI, data = dir.mat)
# compare models, find if interaction term improves model
anova(model.1, model.2)## Analysis of Deviance Table
##
## Model 1: DirichReg(formula = CellTypes ~ COHORT + SEX + AGE + BMI, data =
## dir.mat)
## Model 2: DirichReg(formula = CellTypes ~ SEX + AGE + BMI, data = dir.mat)
##
## Deviance N. par Difference df Pr(>Chi)
## Model 1 -2238.7 20
## Model 2 -1905.9 16 332.75 4 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# Pr(>Chi) is highly significant (p = 0.005655), which means there's strong evidence against the null hypothesis (the simple model is better). model.1 provides a significantly better fit to the data than model.2.
# this implies that the effect of the treatment may be different for different genotypes.
summary(model.1)## Call:
## DirichReg(formula = CellTypes ~ COHORT + SEX + AGE + BMI, data = dir.mat)
##
## Standardized Residuals:
## Min 1Q Median 3Q Max
## Pericyte -2.3381 -0.5866 -0.1134 0.4426 3.3515
## `Endothelial III` -2.1743 -0.5560 -0.0674 0.3703 2.8251
## `Cardiomyocyte II` -3.5172 -0.5003 0.1716 0.9586 3.8810
## Macrophage -0.8989 -0.6024 -0.5169 -0.2391 5.5009
##
## ------------------------------------------------------------------
## Beta-Coefficients for variable no. 1: Pericyte
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.310109 0.478964 2.735 0.00623 **
## COHORTPostmortem -0.842316 0.132392 -6.362 1.99e-10 ***
## SEX2 0.311654 0.134700 2.314 0.02069 *
## AGE -0.002097 0.005574 -0.376 0.70672
## BMI 0.046832 0.015766 2.970 0.00297 **
## ------------------------------------------------------------------
## Beta-Coefficients for variable no. 2: `Endothelial III`
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.400221 0.469958 2.979 0.00289 **
## COHORTPostmortem -0.949205 0.133141 -7.129 1.01e-12 ***
## SEX2 0.288504 0.136175 2.119 0.03412 *
## AGE -0.002895 0.005709 -0.507 0.61207
## BMI 0.045279 0.015923 2.844 0.00446 **
## ------------------------------------------------------------------
## Beta-Coefficients for variable no. 3: `Cardiomyocyte II`
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.9191354 0.4744571 1.937 0.052716 .
## COHORTPostmortem -2.2353731 0.1391499 -16.064 < 2e-16 ***
## SEX2 0.2330513 0.1354878 1.720 0.085416 .
## AGE -0.0008562 0.0060068 -0.143 0.886657
## BMI 0.0574159 0.0161439 3.557 0.000376 ***
## ------------------------------------------------------------------
## Beta-Coefficients for variable no. 4: Macrophage
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.260077 0.560679 0.464 0.643
## COHORTPostmortem -0.846694 0.136261 -6.214 5.17e-10 ***
## SEX2 0.169105 0.142243 1.189 0.235
## AGE -0.008941 0.005752 -1.554 0.120
## BMI -0.003498 0.017680 -0.198 0.843
## ------------------------------------------------------------------
## Significance codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Log-likelihood: 1119 on 20 df (136 BFGS + 2 NR Iterations)
## AIC: -2199, BIC: -2132
## Number of Observations: 207
## Link: Log
## Parametrization: common